WO2010055205A1

WO2010055205A1 - Method, system and computer program for monitoring a person

Info

Publication number: WO2010055205A1
Application number: PCT/FI2009/050905
Authority: WO
Inventors: Reijo Kortesalmi
Original assignee: Reijo Kortesalmi
Priority date: 2008-11-11
Filing date: 2009-11-11
Publication date: 2010-05-20

Abstract

This document presents a method for monitoring a person, a system implementing the method, and an associated computer program product. In the method, images of a monitored space produced by a camera system disposed in the monitored space are utilized. Camera images are analyzed with signal processing equipment by searching from the stored images an object to be monitored and by comparing parameters describing forms and/or motions (gestures) of the object with stored reference parameters corresponding to reference forms and/or reference motions. As a result of an event detected in this pattern recognition analysis, a predetermined action associated with the event is conducted to notify a system operator of the event. This event may be, for example, a medical emergency or a call for help. The system uses cameras that work in any lighting conditions and may use microphones to collect data for analysis.

Description

Method, System and Computer Program for Monitoring a Person

Field

The invention relates to a method and system that are used for monitoring a person or persons and thus it is applicable for overseeing a patient in homecare or in institutionalized nursing or another private person at home.

Background

It has been foreseen that an average age will rise strongly in future years and that the number of people over 65 years will increase. This poses significant challenges to the state, cities and private institutions with respect to nursing elderly people. An object is that as many elderly people as possible is able live independently at home and in familiar environment for as long as possible. This would be both cost-efficient and sensible for the elderly. Prior art proposes numerous methods for monitoring motions and functions of a patient which methods enable homecare of an aged or sick person. Devices attached to the patient's body that monitor patient's vital functions, motions, activity, etc. and that conduct an alarm upon detecting anomalies in the functions of the patient, are in commercial use. A problem with these devices attached to the body of the patient is that they complicate the patient's normal life (e.g. washing) and that the patient may remove such a device. This deteriorates the reliability of monitoring the patient. These devices are generally based on an alarm conducted by the patient himself, which is unreliable, because the patient may not be the right person to make the decision of the necessity of the alarm, and the patient may not know how to make the alarm, or does not want to make the alarm even though it would be necessary.

Motion sensors or video cameras have been proposed to be placed at the patient's home or another place for monitoring motions of the patient in the monitored space. A problem with the motion sensors is that they, when used alone, enable only indicating in which part of the monitored space the patient has moved, and they do not provide a reliable picture of the patient's postures or motions. They cannot indicate whether the patient has fallen, for example. With video cameras, this type of information can be detected when processing a video image with suitable video processing equipment, but recognizing a person from the video image is difficult, because the body of the person may be mixed into the background. Furthermore, the video image is very susceptible to lighting, and monitoring the person's motions at night time is not possible.

Brief description

An object of the present invention is realize a method and a system implementing the method such that monitoring the patient can be carried out reliably.

The invention relates to a method according to claim 1. The invention also relates to systems according to claims 15 and

30, arranged to carry out the method according to claim 1.

The invention also relates to a computer program product according to claim 29 arranged to be executed in the system according to claim 15.

Embodiments of the invention are defined in the dependent claims. With the method and system according to the invention, several advantages are achieved. The invention utilizes a camera system in monitoring the patient. The camera system may be a thermal camera system or another camera system intended for use under good or bad lighting conditions. The quality of an image provided particularly by a thermal camera is completely independent of the lighting, and the person can be easily detected from the image because of high temperature resolution of the thermal image and a technique based on passive imaging. Then, the person is clearly distinguished from the background, which facilitates detection of motions and postures, and the requirement for processing capacity is decreased. The patient does not have to wear any specific device either, which improves the reliability of the system. The system is able to conduct an alarm independently on the basis of an event detected, and alarms can be categorized on the basis of events detected. Monitoring based on the thermal images may also be combined with a fire and burglary alarm, stove watch or other such watches that improve home security. The system is additionally movable and configurable to any location (apartment or institution) easily and affordably. The invention also enables statistical tracing of the patient's activity and habits, which assists in nursing the patient. Additionally, the invention allows detecting a behavior pattern typical to a certain illness which expedites starting the treatment and recovery from the illness. List of drawings

The present invention is now described in more detail in connection with preferred embodiments with reference to the accompanying drawings, in which Figure 1 illustrates a system according to an embodiment of the invention as installed into an apartment;

Figure 2 illustrates a block diagram of a system according to an embodiment of the invention;

Figures 3A to 3C illustrate an event related to a thermal camera image being analyzed according to an embodiment of the invention;

Figure 4 illustrates a thermal camera image filtered according to an embodiment of the invention;

Figure 5 illustrates an event related to a thermal camera image being analyzed according to an embodiment of the invention; Figures 6A and 6B illustrate an event related to a thermal camera image being analyzed according to an embodiment of the invention;

Figure 7 is a flow diagram illustrating a process for monitoring a patient according to an embodiment of the invention;

Figure 8 is a flow diagram illustrating a process for detecting falling of a patient being monitored according to an embodiment of the invention; and

Figure 9 is a flow diagram illustrating a process for detecting illnesses of a patient monitored according to an embodiment of the invention.

Description of embodiments

A system for monitoring a patient according to the present invention comprises one or more cameras storing an image from which a person can be detected very accurately regardless of the lighting conditions. As mentioned above, regular video cameras do not work for this purpose in the dark.

Common night vision cameras based on an infrared light source have a similar problem although they function in the dark. The problem with the night vision cameras based on the infrared light source is that, in addition to the body of a human, the surroundings reflect the radiation emitted by the light source and, thus, separating the human from an image stored by such a night vision device may be very difficult, which degrades the reliability of such a system.

Instead, a thermal camera is inherently passive and based on storing thermal radiation emitted by the object itself and, thus, its separation capability is significantly better than that of a device based on an infrared light source, because the infrared light related to the surroundings does not degrade the separation capability. In a typical indoor environment, the surroundings reflect thermal radiation corresponding to the room temperature and, thus a human body having a temperature of 37 degrees (Celsius) is easily distinguishable from the surroundings as a clearly warmer object. Pets and different pieces of electrical equipment also emit radiation which is warmer than the room temperature, but a human body is distinguishable from these by using signal processing based on the form of the human body. Separating a human body from an image with signal processing based on machine vision is known as such in the art. Recognizing postures of a human detected in an image is known as such in the art of machine vision. Typically, thermal image is processed with signal processing equipment by searching a human figure from a thermal image stored and by comparing parameters describing the postures of the detected human figure to the stored reference parameters of corresponding postures. The human figure refers to the human detected in the thermal image. The human figure is detected from the thermal image, but the intention is not necessarily to identify the person to maintain privacy. When the object is a patient in homecare, the recognition of an identity carried out by the system according to embodiments of the invention is not necessarily even required. If there are several persons in the monitored space whose movement in the monitored space the system is configured to monitor, the system may then assign a virtual identifier to each figure (e.g. patient #1 , patient #2, etc.). Thus, the system may monitor the motions of each patient without needing to identify the identity of the patients. The thermal image produced by the thermal camera may, depending on the embodiment, be a still image, a series of still images, or a video image. The implementation depends on how accurately the motions of the object need to be monitored, on the capacity of the signal processing equipment in the system, etc. In an embodiment of the invention, the system includes a camera system which allows imaging the monitored space and recognition of a human figure in an image under all lighting conditions. The camera system may provide an image or images that enable construction of a three-dimensional model of the detected human figure. The three dimensional model of the object may be constructed by using techniques known as such in the art. In the creation of three-dimensional models, it is preferable to use such imaging devices that can produce images in the gloom or in the dark. A camera system intended to create the three-dimensional model may comprise two or more thermal cameras mentioned above, adapted to capture an image of the monitored space from different angles. Thus, the three-dimensional model of a human figure detected in the images produced by the thermal images may be constructed by using signal processing methods known as such. Alternatively, one active camera may be used in the construction of the three-dimensional model, wherein the camera transmits an infrared beam, measures intensity and delay of a beam reflected from an object and constructs a three- dimensional model of the monitored space on the basis of them. An example of such a camera is described in patent publication WO 0249367 (3DV Systems ltd.). In this application, the system utilizes an active infrared light source and reflection of the infrared light from the monitored space, but the distance of the monitored object can be determined and, thus, the desired object may be separated from the background by using the distance of the object from the camera as a discriminating parameter.

In the following description, thermal camera images are used as an example, but it should be noted that the thermal image used in the analysis may be either a two-dimensional or a three-dimensional model of the monitored space and that the image may be captured with an imaging device other than the thermal camera. An imaging device detecting thermal radiation may, however, be regarded as a preferred embodiment.

Figure 1 illustrates a system according to an embodiment of the invention installed in an apartment for monitoring a patient. The apartment may be the home of the patient and, as a consequence, the system allows homecare of the patient, which is more affordable than institutionalized care and sensible for the patient. The patient may be an aged person or a person suffering from dementia or another psychical or physical disorder and, thus, in the need of regular nursing. The invention is also applicable to private use for improving home security, for example.

In this system, a plurality of thermal cameras 100 to 108 is installed in an apartment such that, with them, the most of the apartment area and, particularly, desired rooms or objects can be monitored. For example, one thermal camera 106 is directed to produce thermal images of the kitchen and, particularly, of a stove. Thus, it can be used to detect if the stove has been left on and an alarm can be conducted, if necessary, as will be described later. There is no need to cover the whole apartment with thermal cameras. It may be desired to not place a thermal camera in the bathroom, for example, in order to maintain privacy, even though the identity of persons is not attempted. Additionally, signal processing equipment is arranged in the apartment to analyze thermal images produced by the thermal cameras by searching a human figure from the thermal image and by analyzing the postures of the human figure detected in the image and/or motions in the apartment. The signal processing equipment may be placed in the same room, apartment or building where the patient is intended to be monitored. Alternatively, thermal images may be transmitted over a telecommunication network to a distant place where the analysis of the thermal images is carried out. The distant place may reside in a different building than the monitored space, for example. The system according to this embodiment of the invention also includes telecommunication means for conducting the possible alarm if the image analysis detects an event that requires alarming a nurse to nurse the patient (or another type of an alarm).

Even though the examples described herein are related to using the monitoring system indoors, the system may be equally applied to outdoors usage. Figure 2 illustrates a block diagram of a patient monitoring system using thermal cameras that may be arranged in an apartment as illustrated in Figure 1 . As mentioned above, the system comprises one or more thermal cameras 100 to 1 10, signal processing equipment 202 and telecommunication means 200. The telecommunication means may include hardware and software necessary for establishing a telecommunications connection to a service provider responsible for nursing and surveillance of the patient. The telecommunications connection may be established with a local area network (LAN) connection to a local area network of the building, with xDSL (Digital Subscriber Line) connection to wirelessly utilizing mobile phone network (GSM, UMTS, WiMAX), for example. The signal processing equipment 202 and the thermal cameras 100 to 1 10 may communicate with each other over a LAN connection (e.g. Ethernet).

The signal processing equipment 202 is included in a control unit 204 of the system and is configured to receive a thermal image of the monitored space from the thermal cameras 100 to 1 10 and to analyze the thermal image by searching a human figure from the thermal image and by comparing parameters describing postures of the person detected in the thermal image with stored parameters corresponding to reference postures and/or motions in order to interpret the motions and postures of the human figure. The parameters corresponding to the reference postures may be stored in a memory unit 206 which may contain parameters for standing, sitting, falling, different signaling (e.g. hand waving) etc. If the signal processing unit 202 detects a correspondence between the posture of a person detected in an image stored by a thermal camera and one of the reference postures, the signal processing equipment 202 provides the control unit 204 with a message about the detection of said posture. The control unit 204 then carries out, in response to the event detected in the analysis, a predetermined action associated with the event in question, which action may be an alarm and/or storing the event in the memory unit. The control unit 204 may be configured to carry out a different action depending on the detected event. For example, if it has been detected that the patient has fallen but rises up, the control unit may store the detection of the falling and the time of the event (date and time) in the memory unit. In this case, the control unit may be configured to notify on the falling a control center connected to the system through the telecommunication means so that a nurse may in his/her next routine visit check whether or not the patient has been hurt. In other words, the control unit 204 may conduct a low priority alarm which requires no immediate measures. On the other hand, if it has been detected that the patient lies on the floor and is not moving (unconsciousness due to falling, for example), the control unit 204 may be configured to conduct a high priority alarm which requires an immediate measure from a nurse or someone else so as to check the situation of the patient. The action performed by the control unit 204 for each event detected on the basis of the analysis may be parameterized in a desired manner. The control unit 204 may include in each alarm a message which indicates the detected event in the form of text or an image. The control unit 204 may be configured to store all detected events to a memory unit 206 together with the time of the event, which allows monitoring the actions of the patient statistically in the long time interval. For example, the signal processing equipment 202 may be arranged to notify the detected event if the patient is detected to walk in the middle of the night (or another determined time of the day). Thus, the amount of sleep of the patient may be monitored. The system according to Figure 2 may be realized by one or more what are called intelligent cameras. An intelligent camera is referred to as equipment which includes a camera (thermal camera in this case) and a processing unit configured by software to process images produced by the camera. Referring to Figure 2, a system according to an embodiment of the invention may be realized by one or more intelligent cameras, wherein each intelligent camera includes a thermal camera, signal processing equipment 202, a control unit 204, a memory unit 206 and telecommunication means in a single entity (e.g. in a single casing). The memory unit 206 may be a memory card which is detachably attached to the intelligent camera and adapted to store software and parameters configuring the processing unit. When constructing a three-dimensional model of the human figure, one unit of the signal processing equipment of the intelligent cameras (or additional external signal processing equipment) may construct the three-dimensional model in a centralized manner on the basis of the received images and, then, analyze the three-dimensional model by comparing it with the stored reference parameters. Each one of the intelligent cameras may first separate the human figure from the thermal image and then send the image of the human figure to the signal processing equipment constructing the three-dimensional model. The human figure may be separated from the thermal image by filtering out objects other than the human figure, as will be described below, or by separating the human figure on the basis of extrinsic parameters, such as measures (width, height, form).

Figures 3A to 3C illustrate thermal images analyzed according to an embodiment of the invention, wherein the images represent an event related to a patient falling. The image includes a human figure 302 corresponding to the patient and a lamp 304 belonging to the background. Only the part of the lamp 304 that radiates heat is shown. The signal processing equipment 202 searches the image for a human figure on the basis of stored form and/or temperature parameters related to the human figure. A typical machine vision algorithm is based on recognizing a human head from the thermal image. Upon detecting the human figure, the signal processing equipment starts to analyze the motions and/or postures of the human figure by comparing the detected postures with reference postures or reference motion vectors. When the signal processing equipment 202 detects that the postures and/or motion vectors of the human figure correlate sufficiently (similarity/correlation exceeds a determined threshold level) with reference postures and/or motion vectors associated with falling, the signal processing equipment 202 determines that the patient has fallen and notifies the control unit 204 about an event associated with the patient falling. Then, the control unit 204 initiates a process associated with the event of patient falling.

As mentioned above, the signal processing equipment 202 is able to discriminate the human figure from the thermal image on the basis of the shape and/or temperature information. The signal processing equipment 202 analyzes the pixels of the thermal image, wherein each pixel has a value describing the temperature. Since the temperature of the human body is significantly warmer (around 37°C) than the surrounding air or passive items (furniture, plants etc.) in the room, the signal processing equipment 202 may be adapted to filter out temperatures that deviate from a selected temperature range corresponding to the human body temperature in order to discriminate the human figure from the background. The signal processing equipment 202 may filter out from the image temperatures lower than 35-40⁰C when analyzing the motions and postures of the human figure. The signal processing equipment 202 may also filter out from the image temperatures higher than 35- 40⁰C so that for example a stove, a television, and other electronic devices radiating heat would not interfere with the analysis. After such filtering, objects other than the human body and other objects radiating heat in that temperature range are removed from the image, as illustrated in figure 4 with respect to Figures 3A to 3C. This reduces the amount of information in the image being processed, which reduces the processing capacity required from the signal processing equipment and which is well applicable to an embodiment where the image to be processed is transmitted over the telecommunication connection to a distant place (reduces the required data transfer capacity). Then, the signal processing equipment may separate the human body from an animal, for example, on the basis of shape analysis. In a process monitoring the environment and carried out by the signal processing equipment 202 that can be executed in parallel with motion analysis of the human figure, the signal processing equipment 202 analyzes an unfiltered image in order to monitor conditions in the apartment, or it even filters the human body out of the image by filtering out the pixels corresponding to the human body temperature. The signal processing equipment 202 may, for example, monitor an average temperature in the apartment and, if the temperature is detected to be below an acceptable minimum level (as a sign of poor heating) or over an allowed maximum level (e.g. fire), the signal processing equipment 202 provides the control unit 204 with a signal about the corresponding event. Upon reception of the fire event, the control unit 204 may be arranged to conduct a high level alarm, and upon reception of too low temperature, the control unit 204 may be arranged to conduct a low priority alarm and notify information about said event, on the basis of which the nurse knows that he/she has to check the room temperature in his/her next routine visit. The signal processing equipment 202 may also be configured to recognize a patient's signaling on the basis of the motion and posture analysis. The signal processing equipment may be configured to detect a request for help by the patient when comparing the motions of the human figure detected in the thermal image with stored reference motions associated with the request for help. The signal processing equipment 202 may associate the detected motion with an explicit call for help. Then, the signal processing equipment 202 may notify the control unit 204 about the request for help event, whereby the control unit 204 carries out a process associated with the request for help, which may include conducting a predetermined alarm associated with the request for help. The control unit may additionally notify the patient through an audible signal or by other signaling means about conducting the alarm. The alarm is preferably a high priority alarm requiring immediate measures from the nursing side. The request for help may be defined beforehand to be conducted by waving a hand from left to right three times. As a consequence, the reference parameters corresponding to said trajectory may be stored beforehand in the memory unit 206. Of course, another motion may be defined as a sign for requesting the help but, preferably, a trajectory not made accidentally by a human is selected as the request for help. This minimizes the number of false alarms. As a precaution for a false alarm, the signal processing equipment may also be configured to detect a motion associated with cancelling the alarm. If the patient discovers on the basis of signaling provided by the control unit that he/she has made a false alarm, he/she may cancel the alarm with the predetermined signaling.

The motion and/or posture associated with the call for help as described above enables the patient to conduct an alarm intentionally. As mentioned above, the signal processing equipment associates the corresponding posture and/or motion with explicit call for help causing the execution of an immediate alarm. With respect to other events causing the alarm, the signal processing equipment detects the event causing the alarm by analyzing the motions and/or postures (and/or voice signals) of the patient for a given duration and determines the need for the alarm implicitly, i.e. the patient does not cause the alarm intentionally but through motions and/or postures unrelated directly to causing the alarm intentionally. In such implicit determination of whether or not the alarm is made, it may take a while for the signal processing equipment 202 to detect the event but, on the contrary, with explicit request for help enabling the patient to call for help intentionally, the alarm may be conducted more rapidly.

Figures 6A and 6B illustrate thermal images related to an embodiment of the invention, to which images the signal processing equipment 202 is adapted to define one or more special zones 600 associated with one or more special operational models. The special zone may be a portion defined in the thermal image. The operational models are stored in the memory unit 206 beforehand, from which the signal processing equipment 202 may read the operational models when detecting a predetermined event in the special zone triggering the utilization of the operational models. The predetermined event may be, for example, detecting a human figure 302 entering the special zone 600 or a rise in the temperature in the special zone (e.g. the stove). The operation models may define events allowed in the special zone, whereby no alarm is necessary. If the signal processing equipment 202 detects motions of the human figure not corresponding to the reference parameters associated with the allowed events, the signal processing equipment 202 provides the control unit 204 with a signal about an event in the special zone deviating from the allowed events. The control unit 204 may then carry out an action, such as conducting an alarm through the telecommunication means 200 and/or storing the event in the memory unit 206. Figures 6A and 6B illustrate an example of this embodiment. In this example, the surveillance system according to the invention is located in a nursing home or a hospital, for example, and the special zone 600 is a lounge of the patients. Let us assume that it is not desired that the patients leave the lounge by themselves without the guidance of a nurse. Then, utilization of the allowed operational models may be triggered by a human figure 302 detected by the signal processing means to enter the special zone 600 or staying in the special zone 600 for longer than a predetermined duration (if the patient is escorted to the zone). The signal processing equipment may count the predetermined time elapsing by starting a timer upon detecting the human figure entering the zone 600 and utilize the operational models if the human figure is not detected to leave the zone before the timer has expired.

When the signal processing equipment 202 detects the utilization of the allowed operational models when the human figure enters the special zone 600, it reads from the memory unit 206 parameters corresponding to the allowed operational models. An allowed operational model stored in the memory unit 206 may, for example, include motion parameters for that another person 602 entering the special zone 600 (Figure 6B) to escort the patient 302 away from the special zone. Escorting may be interpreted as human figures touching one another while leaving the zone 600. The touch shows in the thermal image clearly so that pixels corresponding to the human figures are in connection to each other without 'cold' pixels belonging to the background between the two human figures. In order to improve the reliability of the image analysis in this case, multiple thermal cameras may be arranged to monitor the special zone. Then it can be distinguished whether one person is escorting another person out of the special zone or whether they are moving one behind another. One thermal camera imaging from a certain view angle does not necessarily provide this information. Therefore, the signal processing equipment 202 may analyze multiple thermal images from the same special zone in order to detect on the basis of every thermal image, whether the patient is escorted by another human figure in the manner described above. If the signal processing equipment 202 detects the first human figure 302 exiting the special zone in connection with another human figure 602 that has entered the zone 602 (in every thermal image), this is seen as an allowed operation model and no alarm is conducted. However, if the signal processing equipment 202 detects the first human figure 302 exiting the special zone alone, this is seen as an operation model deviating from the allowed reference, and the signal processing equipment 202 provides the control unit 204 with a signal about a prohibited event in the special zone (identification information of the special zone may also be provided), whereby the control unit 204 may conduct an alarm. The allowed operational model may include other parameters to prevent that patients do not exit the special zone in pairs, unless exiting in pairs is allowed. Assuming that a patient is allowed to exit the lounge alone, if he/she is in a conscious state, the system may allow the patient to exit the zone without an alarm if the patient shows by an agreed signal (e.g. waving a hand) that he/she is in a conscious state. Then, if the signal processing system detects the predetermined signal before the human figure exits the special zone, the exiting is seen as an allowed event and no alarm is conducted.

For example, a couch or a stove may be set as corresponding special zones in the patient's apartment. Let us assume that it would not be desirable for the patient to watch television on the couch more than a determined number of hours. When the signal processing equipment 202 detects a human figure moving onto a special zone including the couch (the human figure is detected inside the special zone), a timer is started. When the signal processing equipment 202 detects that the patient leaves the special zone (the human figure is no longer detected in the special zone), the timer is stopped and the time measured by the timer is stored in the memory unit. The control unit 204 may monitor the stored timer recordings in units of a day. If the control unit 204 detects that the measured time shown by the timer recordings exceeds a determined limit, the control unit 204 stores in the memory unit 206 information on watching the television over the allowed limit in the certain day. In another embodiment, the allowed operation models associated with the special zone(s) include models to be used when entering the special zone. Accordingly, the signal processing equipment may monitor the manner in which the human figures enter the special zone and, if a human figure enters the special zone in a prohibited manner, an alarm is conducted. As already disclosed above in connection with exiting the special zones, the system may allow a person to enter the special zone after indicating a determined motion and/or posture to the system. The signal processing equipment 202 may analyze the motion and/or posture and determine whether or not it correlates with a reference motion and/or posture associated with allowed entry to the special zone. Analogously, a human figure may be allowed to enter the zone in connection with another human figure or human figures. In another embodiment, a human figure is allowed to enter the special zone without conducting the alarm only at defined times of the day. For example, no human figure may be allowed to enter through a front door during night time. The signal processing equipment may also be arranged to monitor the activity of the human figure in the special zone. The activity may be determined by analyzing the posture of the human figure (sitting, lying, standing) and/or the degree of motion of the human figure. The activity may be monitored in the special zone around the clock (monitoring is performed only when the human figure is detected in the special zone) or during determined periods of time. A certain bedtime may be assigned to the patient, when the patient has to move to the bed. Then, the signal processing equipment 202 may be arranged to monitor the patient in the special zone (couch in this example) from the time related to the bedtime onwards. If the signal processing equipment 202 detects the human figure in the special zone, the signal processing equipment may provide the control unit 204 with a signal about the event, and the control unit 204 may transmit through the telecommunication means a message about the event to an overseer (low priority alarm). Then, the overseer is aware to call the patient, for example, and knows he/she should tell him/her to go to bed. The signal processing equipment 202 may additionally be arranged to monitor the posture and/or activity of the human figure in the special zone. If it has been detected that the human figure is lying and moves infrequently, the signal processing equipment 202 may decide that the patient has fallen asleep on the couch and, thus, provides the control unit a signal about the event. Then, the control unit 204 may send a message about the event to the overseer through the telecommunication means (low priority alarm). In general, the activity of the human figure may naturally be monitored in other parts of the monitored area than the special zones either around the clock or during determined time periods. The stove can be set as the special zone so that the system may monitor the stove being left on and warn about the fire in time. In this case, the utilization of the allowed operational models may be triggered by the signal processing means 202 detecting the rise in the temperature above a predetermined threshold in the special area where the stove resides. The threshold may be, for example, 100⁰C or even a higher temperature indicating that the stove is on. The temperature is preferably such that it can be distinguished from the thermal image when no pot or such is on the stove. As a consequence, long-term cooking does not cause the alarm. When the signal processing equipment 202 detects the temperature above the threshold, it starts a timer. If the temperature does not fall below the threshold before the duration measured by the timer expires, the signal processing equipment notifies the control unit about an event related to the stove being left on. The control unit 204 may then be configured to notify in the apartment with a certain audible signal about the stove being left on and start a timer. The signal processing equipment 202 may continue monitoring the stove and notify the control unit 204 if the temperature is detected to fall below the threshold. If the control unit 204 does not receive a notification about the stove being switched off from the signal processing equipment within a determined time measured by the timer, the control unit 204 conducts a high priority alarm through the telecommunication means 200. Figures 7 and 8 illustrate processes to implement the method according to the invention. The processes may be carried out by computer software with which operation of one or more processing units may be controlled. Thus, the software may be physically distributed into multiple entities. The signal processing means 202 and the control unit 204 of Figure 2 may be realized by one or more processors controlled with the computer software. The software may be stored on a computer readable record media, such as a memory chip, hard disk, optical record medium, etc.

Figure 7 illustrates on a general level a process for monitoring a patient, and Figure 8 illustrates a process for detecting the falling patient for performing necessary action by the system. Referring to Figure 7, the process is started in block 700. The process may be started with the reception of thermal images (video) from a camera system at the signal processing means 202 or with a command from the control unit 204 to start the process. In step 702, a human figure is searched from the received thermal image or images. The search may be conducted by searching from the image pixels corresponding to the human body temperature. Additionally, the shape of an object formed by the pixels corresponding to the human body temperature may be analyzed by comparing the shape to the stored human reference shape (one or more). The search may be facilitated by executing step 704, i.e. by separating the human figure from the image, e.g. by filtering out of the thermal image one or more temperature ranges that do not correspond to the human body temperature. Separation based on other parameters typical of a human figure (such as extrinsic measures) is a possible embodiment. Step 704 is optional, which means that it may be omitted from the process. When the signal processing equipment 202 has detected a human figure in the thermal image, it constructs in step 705 a three-dimensional model of the human figure from the thermal image or images and compares the postures of the three-dimensional model to the stored reference parameters in order to determine the motions and postures of the human figure in step 706, as described above. Step 705 is optional, and the process may be carried out by using two-dimensional images. In step 708, it is determined whether an event causing an alarm is detected in the motions or postures of the human figure. This step may be carried out in the signal processing equipment 202 with respect to the detection of the event on the basis of the analysis and in the control unit 204 with respect to the decision of the necessity of the alarm. If in step 708 it is decided that either no event has been detected or that the detected event does not cause the alarm to be conducted, the process returns to step 706. If an event has been detected, the control unit may store information on the event when proceeding from step 708 to step 706. If in step 708 it is determined that an event has been detected as a result of which the alarm needs to be conducted, the process proceeds to step 710 where the alarm is conducted. Conducting the alarm may include determining the manner of conducting the alarm associated with the event before conducting the actual alarm. This allows conducting different types of alarms on the basis of the detected event. Each event, whether it is related to the human figure or to another object being monitored, may be associated with one of a plurality of different alarm types, e.g. different priorities, and the control unit 204 may conduct the alarm according to a protocol associated with the alarm type. For example, the detection of the patient falling does not necessarily cause the high priority alarm, as described in the following example. The process returns from step 710 to step 706 where monitoring the motions of the patient is continued.

Figure 8 illustrates a process related to the detection of the patient falling. Step 800 is entered from block 705 after the three-dimensional model has been constructed. In step 800, the signal processing equipment is configured to compare the postures of the human figure modeled by the three- dimensional model with one or more reference parameters associated with the falling. In step 802, it is determined on the basis of step 800 whether or not the motion of the detected human figure and the reference parameters associated with the falling correlate with each other sufficiently, i.e. whether or not the falling has been detected. If the falling has not been detected, the process returns to step 800. If the falling human figure has been detected, the process proceeds to step 804 where a tinner is started. In step 806, monitoring the motions of the human figure is started in the signal processing equipment 202. In step 808, it is determined whether or not the human figure moves sufficiently. If it is determined that the human figure moves sufficiently, the process returns to step 800. In the other case, the process proceeds to step 710. The motions may be detected with the pixels representing the human body temperature switching, and a limit for the switching may be set for the detection of the sufficient motion, for example.

Above-mentioned embodiments describing the operation of the signal processing equipment 202 and the control unit 204 may naturally be combined to realize new embodiments. For example, the embodiment describing the process of Figure 8 may be combined with the embodiment of Figure 5 related to the request for help. Let us assume a situation where the patient falls and hurts his/her leg. Then, he/she is able to move so the process of Figure 8 does not necessarily conduct the alarm. Then, the patient may provide the signal agreed to be the request for help, whereby a process for detecting the request for help signal, for example through the functionality described in connection with Figure 5 and run in parallel, detects the request for help and conducts a high priority alarm. Alternatively (or additionally), the process of Figure 8 may be improved to enable detection of the patient injured as a result of falling down, for example. In this embodiment, in step 806 it is monitored whether or not the human figure rises up to stand or sit within the time interval measured by the timer. Here, the time interval measured by the timer may be set to be sufficiently long so that it exceeds the time duration of temporary unconsciousness caused by the falling. On the basis of the above description, the signal processing equipment 202 is able to distinguish on the basis of the analysis and the reference parameters whether the human figure has been detected in a lying, sitting, or standing posture. In step 808 it is determined whether the human figure is detected to move into the standing or sitting posture. If this is detected, the process returns to step 800. Otherwise, the process moves to step 710.

As another example of combining the embodiments, it is described how the embodiment monitoring the surroundings may be combined with the embodiment using the special zones. With this embodiment, for example, a burglar may be detected. The signal processing equipment 202 may be adapted to set as special zones in the thermal image or images locations corresponding to anticipated entry routes of the burglar, such as the outer door and/or windows. The signal processing equipment 202 may be configured to carry out monitoring these special zones at a certain time of the day (at night time, for example). When the signal processing equipment 202 detects motions of a human figure in these special zones, it provides the control unit 204 with a message indicating burglary, whereby the control unit 204 conducts an alarm. As another embodiment for detecting the burglar, the signal processing equipment 202 may be configured to monitor the number of human figures in the monitored space at a certain time of the day (e.g. at night) and to compare the number of detected human figures to a predetermined allowed number. If the number of detected human figures exceeds the allowed number, the signal processing equipment 202 notifies the control unit about the corresponding event, and the control unit 204 conducts a corresponding alarm. The signal processing equipment may also store the number of detected human figures in the memory unit so that a record of the number of people moved in the monitored space may be kept. In this embodiment, the recognition based on the human form is used so that the electronic devices or animals are not interpreted as humans. In another embodiment, the signal processing equipment 202 is configured to detect whether an assault or a fight is being conducted in the monitored space. If a person enters the apartment through the front door in the daytime, the system does not necessarily detect the person as a burglar or another person not allowed to enter the apartment. In a typical assault situation, the thermal image includes a plurality of human figures in the same image, the human figures are physically connected to each other, the movement of the human figures is rapid, and/or at least one person falls down and/or remains lying on the ground. The signal processing equipment 202 may be configured to monitor occurrence of such properties in the thermal image and, if sufficient correlation with reference parameters associated with an assault (or fight) is detected, the signal processing equipment 202 notifies the control unit 204 of the assault in the apartment, and the control unit conducts an alarm of a determined alarm type by using the telecommunication unit 200. The alarm type may be, for example, prioritized to a given priority level (by using a numerical priority value, e.g. Priority 1 , highest priority), and/or a protocol including transmission of message "Assault". When generalizing this concept, another alarm type may be used in connection with another event, wherein the alarm type may be prioritized to a given priority level (by using a numerical priority value, e.g. Priority X, X^th highest priority), and/or a protocol defining action(s) to be made when conducting the alarm. In another embodiment, the system is equipped with one or more microphones disposed in the monitored space, and the signal processing equipment 202 is configured with voice recognition means so as to analyze voice (and/or other audio) signals received from the one or more microphones. The voice recognition means may be realized in a manner known in the art, e.g. with a processor configured by suitable software and a memory unit storing the software and reference parameters for different voice patterns the voice recognition means are configured to recognize.

Let us now consider a few examples to which the voice recognition may be applied. In an embodiment, the voice recognition means are configured to detect a voice signal cancelling an alarm that has been conducted. A given voice signal pattern is associated with canceling the alarm, and the voice recognition means may be configured to monitor that voice signal pattern after the alarm has been triggered. When the voice recognition means detect the voice signal pattern associated with canceling the alarm, the voice recognition means may provide the control unit with a signal indicating that the alarm should be canceled. Then, the control unit may cancel the alarm by transmitting an alarm cancel signal to a system operator through the telecommunication unit 200. This allows for canceling an alarm triggered accidentally. Additionally, or alternatively, the voice recognition means may be used to detect the patient triggering the alarm by a voice signal. The voice recognition means may detect the activation of the alarm upon detection of a voice pattern matching, i.e. having sufficient correlation with, a stored reference signal pattern. The reference signal pattern may comprise a voice signal for word "HELP" as pronounced by the patient being monitored, and it may have been stored in the memory unit in connection with the installment of the system. Naturally, other voices may be detected by comparing the received voice signal with stored voice signals and, upon detection of a given voice signal, an action associated with the detection of said voice signal is carried out. In another embodiment utilizing the special zones, a person may enter or leave a given special zone after providing a certain voice signal. With respect to the embodiment of Figures 6A and 6B, the person may be allowed to leave the special zone (lounge) after providing this voice signal. The voice recognition means are configured to monitor the voice signal associated with leaving the special zone when one or more persons are detected inside the special zone and, when the voice signal is detected, the person is allowed to leave the voice signal without conducting an alarm. On the other hand, if the person leaves the special zone without providing the voice signal or in any other prohibited manner, the alarm is triggered. Obviously, the embodiment allowing a patient to leave the special zone(s) after providing the voice signal may be combined with the embodiment of another person escorting the patient.

In another embodiment, the voice recognition means are configured to associate a certain voice signal or signals with allowed entering to the special zone(s). With respect to a front door as the special zone, a person may be allowed to enter the apartment without causing an alarm, if he/she provides an agreed voice signal, e.g. a name or errand in the apartment (cleaner etc.). If a person enters the apartment without providing the agreed voice signal, the alarm may be triggered, or the system may utilize other measures so as to identify the person entering the house.

The voice recognition means may also be used to detect certain illnesses, e.g. flue or other diseases with which certain types of voices are associated. For example, the flue is typically associated with repeating sneezes, and the sneezes may be recognized as occurring with certain frequency and being rapid, possibly high-intensity voices. The voice recognition means may be configured to monitor determined parameters in the received voice signals, e.g. frequency, signal rise time, intensity, etc., and upon detection of fulfillment of determined conditions it may recognize the voice signal. For example, if the voice recognition means detect a rapid signal with a given frequency with which sneezes occur, the voice recognition means determine that a sneeze has occurred. If a given number of sneezes is detected within a certain duration, i.e. with sufficient frequency of occurrence, the voice recognition means may send a message to the control unit, indicating that the patient potentially has a flue. As a consequence, the control unit may conduct a determined protocol so as to notify the event through the telecommunication unit. Similarly, other illnesses detectable from voice patterns may be detected.

The detection of a certain illness (or syndrome) may also be based on detection of repetition of a certain voice pattern. If the voice recognition means detect frequent occurrence of the same voice pattern within a given time duration, it may notify the control unit and the control unit may notify an operator. For example, certain syndromes, e.g. the Tourette syndrome described below, are associated with repetitive chanting voices that may be repeated within short periods of time, e.g. a few minutes, between the chants. When the voice recognition means detect a voice pattern repeated every few minutes (or within a given time window) and for a given duration, which may be longer than a few minutes, the voice recognition means may notify the control unit of the event, and the control unit may notify the operator of the event that the person is repeating the same voices over and over. This notification may be associated with conducting a specific type of an alarm.

Thermal resolution of the thermal images may be so high that the human temperature may be measured with it. The signal processing system may be configured to measure the temperature of the human figure detected in the thermal image by studying values of the pixels associated with the human figure. The temperature measurement may be focused on such part of the human figure that is known to be exposed, e.g. the head. Then, the clothing does not interfere with the measuring result. If the measured temperature exceeds a threshold level, the signal processing system may provide the control unit with a signal of the event, whereby the control unit may conduct an alarm associated with the high temperature of the patient. Similarly, if the measured temperature is below another threshold level set as a lower limit, the signal processing system may provide the control unit with a signal of the event, whereby the control unit may conduct an alarm associated with too low a temperature of the patient. As mentioned at the beginning, the thermal cameras are not necessarily disposed to cover the whole apartment due to the cost or other reasons. For example, the bathroom may be left uncovered. Then, the signal processing equipment 202 may be configured to start a timer when the human figure is detected leaving a bathroom door defined as a special zone into the bathroom. The signal processing equipment may detect this by detecting that the 'warm' pixels associated with the human figure disappear from the thermal image inside the special zone associated with the bathroom door and that they do not move outside the special zone from the edge of the special zone. If the human figure is not found exiting the bathroom in any thermal image (as appearing in the special zone of the bathroom elsewhere than from the edges of the special zone) within a time interval measured by a timer, the signal processing equipment 202 provides the control unit 204 with a signal of the corresponding event, whereby the control unit 204 may send an alarm to the overseer.

With respect to the previously described embodiment, the signal processing equipment 202 may be configured to set a location corresponding to the front door as a special zone in the monitored image. Additionally, the signal processing equipment 202 may count the number of people inside the apartment. If the signal processing equipment 202 detects that all the people counted in the apartment have exited the front door, the signal processing equipment 202 considers this as an allowed event, so the alarm is not conducted in this case. However, if the patient leaving out of the front door is defined as a prohibited operational model (e.g. at night time), the signal processing equipment 202 provides the control unit with a signal indicating the event of prohibited exiting upon detection that the only human figure in the apartment exits the front door. Then, the control unit 204 may conduct the alarm. The patient may, however, be allowed to exit the front door during daytime and/or at other predefined times.

Let us now consider the different types of alarms mentioned above. As already described, the system may employ different priority levels of the alarms and different manners in which the alarm is conducted, e.g. storing a note in the database and/or sending an alarm signal to a system operator through the telecommunication unit 200. The number of different priority levels naturally depends on the implementation and the number of different protocols the nursing party wants to utilize in response to the alarm. The number of priority levels may be, for example, four when the nursing party has established four different operative protocols as how to respond to alarms. With respect to an event associated with the lowest priority alarm, e.g. the patient is walking in the middle of the night, the control unit 204 may merely store in the database a record indicating the occurrence of said event. The nursing party may regularly check the database for any new events and handle the new events accordingly in connection with regular routines. With respect to an event associated with the second lowest priority alarm, e.g. the patient falling down but standing up, or a low room temperature, the control unit 204 may store in the database a record indicating the occurrence of said event and transmit a notification to the system operator through the telecommunication unit 200 so that the operator is able to detect the new event as soon as it occurs and the event handling can be expedited with respect to the lowest priority level. With respect to an event associated with the second highest priority alarm, e.g. the patient falling down but rising up to a sitting posture, or the stove being left on, the control unit 204 may send an alarm signal to the system operator, and the system operator immediately alerts the nursing staff to check the patient. Additionally, the control unit 204 may store a notification in the database. With respect to an event associated with the highest priority alarm, e.g. fire, patient falling down and not moving, burglary, assault, etc. the control unit 204 may send an alarm signal to the system operator and to another party, e.g. to a general emergency address to alert firefighter, police, ambulance, etc. in addition to the nursing staff. Additionally, the control unit 204 may store a notification in the database. The priority is not necessarily the only feature that characterizes the manner in which the control unit carries out the alarm. Within any one (or more) of the priority levels, a plurality of different protocols for conducting the alarm may be provided, and each event may even be associated with a different protocol as how to conduct the alarm.

It may arise a situation where, during a reboot or another corresponding situation, no human figure is detected in any of the monitored thermal images and the system has not detected the human figure exiting the space. Then, the system may start a timer. If no human figure is detected in of the thermal images within a determined time interval measured by the timer, the signal processing equipment 202 provides the control unit 204 with a signal about the event in question, whereby the control unit 204 may send an alarm to the overseer. The system may include, in addition to the thermal cameras, other sensors with which the object may be monitored. For example, a motion sensor providing the control unit 204 with a signal upon detection of movement in an area covered by the motion sensor may be installed in the bathroom. The control unit 204 may then combine the information provided by the motion sensor with information provided by the signal processing equipment. For example, if the signal processing equipment 202 notifies on an event that the human figure is detected leaving the special zone of the bathroom door for the bathroom and that the patient has not been detected as exiting the bathroom through the bathroom door within the determined time, the control unit 204 may check whether or not the motion sensor in the bathroom reports on movement in the bathroom. If a signal provided by the motion sensor indicates that there is movement in the bathroom, the control unit 204 determines that the patient is in the bathroom and moves in there and that no alarm is conducted. On the other hand, if even the motion sensor has not reported any movement within a determined time, the control unit 204 may conduct the alarm.

In an embodiment of the invention, the system is configured to detect a behavior pattern typical of a certain illness. Figure 9 illustrates a process for detecting a behavior pattern typical of an illness according to an embodiment of the invention. The process starts in step 900. The system detects the behavior pattern by analyzing (702, 706, 908) one thermal image(s) produced by one or more thermal cameras with the signal processing equipment by searching (702) from the stored thermal image a human figure, by optionally constructing a three-dimensional model of the human figure (705), and by comparing (706, 908) parameters representing postures and/or motions of the human figure modeled by the (three-dimensional) model with stored reference parameters of respective reference postures and/or reference motions associated with selected illnesses. The system may, for example, monitor the walking speed of the person over a long time interval so as to detect changes in the walking speed. Additionally, the system may be arranged to detect staggering, tremors and general changes in activity. These behavior patterns and their appearance frequency may be monitored within a determined time window. The length of the time window may be adjustable case-specifically, and a time window of different length may be arranged for each behavior pattern. The time window may be adjusted to be so long that occasional passiveness caused by a fever, for example, does not cause further measures. In this manner, the system may be used to detect chronic illnesses. On the other hand, the time window may be adjusted to be short in order to detect temporary sicknesses. The time window may be a sliding time window having a starting point and an ending point flowing in real time, or the system may carry out the monitoring with the time window periodically such that monitoring related to a new time window is started after ending the monitoring in the previous time window. In both cases, the monitoring does not end when the time window ends, but the monitoring is carried out regularly. Alternatively, the time window may have at least the ending point open, which means that the system does not monitor the occurrence frequency of symptoms related to illnesses as such but their number. Upon the number of certain symptoms exceeds a threshold, the control unit may be configured to conduct an alarm or to notify about the symptoms by other means.

Reference parameters corresponding to behavior patterns typical of different illnesses may be stored, and the motions, postures and/or activity of the patient corresponding to the reference parameters within a certain time window causes the system to decide that the person follows the behavior pattern of a certain illness (step 908). Then, the system conducts a predetermined action as a result of the behavior pattern detected in the analysis so as to notify to a system operator on the information related to the behavior pattern. The action may include registering information related to the behavior pattern in a database and/or conducting an alarm by using the telecommunication unit (200) of the system. The process of Figure 9 may be executed as a computer program which controls a processor reading the computer program to execute the steps of Figure 9. For example, when diagnosing the Parkinson's disease, changes in walking are detected: a reduction in unilateral follow motion when walking (for example one arm is significantly stiffer or even stable and bent when walking), steps becoming lower and a reduction in speed. In about one percent of the cases, tremors at rest position appear, first unilaterally when a limb is at rest. The signal processing equipment 202 may be configured to monitor a momentary walking posture and walking speed over a long term. Parameters describing the person walking may be stored as a reference walking posture and speed in the memory unit 206 when installing the system. Additionally, or alternatively, the system may carry out calibration of the reference parameters corresponding to the walking posture and walking speed of the monitored person after a certain time has elapsed from the installment. Then, the monitored person is probably alone in the monitored space and his/her gestures are not influenced by external factors. The signal processing system may periodically compare the walking posture and the walking speed of the person obtained from the thermal image with the reference parameters within a long time window. The length of the time window may be, for example, in the order of weeks. The equipment 202 may compare the walking speed detected within the time window with a reference speed, and if the detected walking speed is below the reference speed more than an allowed threshold, the system may store in the database a record that the walking speed of the person is slowing down. Additionally, or alternatively, parameters corresponding to a walking posture typical for the Parkinson's disease may be stored in the database, whereby the signal processing system may compare the walking posture of the person with these reference parameters. If the system detects in the comparison that the walking posture of the person repeatedly corresponds to the reference parameters corresponding to the Parkinson's disease, the system may store in the database a record of a possible Parkinson's disease so that an instance maintaining the system knows that they should take the patient into tests for a more accurate diagnosis. Instead of storing the record (or in addition to that), the system may conduct a low priority alarm through the telecommunication unit.

With respect to the MS disease, certain neural damages may cause that the patient is not able to control the functionality of his/her muscles normally. This is called spasticity. As a result, there exist spasms in one or more muscle groups which is related to aptitude for muscle twitches. The human organism reacts to the muscle contraction but cannot affect to it willfully. In order to detect a behavior pattern typical of the MS disease, the system 202 may be adapted to monitor the speed of motions of a person by comparing the speed of motions of the person to a reference threshold adapted to be so high that normal motions of the person do not exceed the threshold. Frequency of occurrence of motion speeds exceeding the threshold may be monitored within a time window long enough so that muscle spasms during sleep, for example, do not cause measures. The time window may be a day or several days, for example. If the frequency of occurrence within the defined time window exceeds the threshold, the system may provide a low priority alarm and/or store in the database a record of a detected behavior pattern related to the MS disease.

The Tourette syndrome is a neurological or neurochemical syndrome having as symptoms tic motions, i.e. abrupt and rapid spasms and motions and voices that are not completely willfully and are repeated in the same manner over short time periods. The detection of a behavior pattern typical for the Tourette syndrome may be realized by the system in the same manner as the MS disease. In this embodiment, the voice recognition means may be utilized to detect the voice patterns associated with the Tourette syndrome. Since the detection of MS and Tourette may be similar, the system does not necessarily store in the database a record of detection of a behavior pattern typical of a certain illness but, for example, a record of the detection of a behavior pattern deviating from a normal one. In addition to the diseases mentioned above, the system may be configured to monitor behavior patterns detectable visually from motions and/or postures of a person and typical of other illnesses. Additionally, or alternatively, audible detection may be utilized in the form of voice recognition. Thus, occurrence of different chronic diseases may be discovered automatically and fast, which expedites start of the treatment and improves the recovery estimate.

The record may be a character-based note of the detected behavior pattern but the system may also store thermal images in a defined format, e.g. video images or still images with sufficient frequency (e.g. 10 images per second). The stored thermal images may facilitate the nursing staff and doctors to make a diagnosis. With the help of the video images (or still images captured with a sufficient frequency, e.g. 10 images per second) stored by the system, the behavior pattern of the person may be monitored in his/her natural environment at home, for example, without the psychological effect of a hospital or the like on the behavior of the person. Upon deciding on the detection of a certain behavior pattern, the control unit 204 may be configured to store video images into the database the next time when a symptom related to said behavior pattern is detected (sudden motion exceeding the threshold, for example). The system may buffer the video images for one minute, for example, whereby video images may be stored from a time period before the detection of said symptom. The video images may include images from motions of the person for a determined duration starting from a time instant before detecting the symptom and ending at a time instant after the detection of said symptom. The length of the video may be 30 seconds, for example.

In addition to the detection of behavior patterns related to illnesses, the effect of medication on the behavior pattern may be monitored with the system. The control unit 204 may store in the database (e.g. memory unit 206) the frequency of occurrence of motions or other such symptoms detectable from the thermal images and associated with a certain illness in the table form, which allows for monitoring the status of the patient from the database. The control unit may be configured to store particularly the frequency of occurrence of symptoms related to a diagnosed illness. Such symptoms may be staggering events of the patient, tremor episodes, dyskinesia classified as abrupt, recovery from unilateral follow motion, etc. The system may also monitor the activity of the patient and store in the database activity with a certain index or an indicator indicating the increase/decrease of the activity.

Above, the operation of the system is described mainly when monitoring the functions, postures, motions and/or voices of a single person. Naturally, the system may be configured to monitor multiple human figures as parallel processes. Although the invention has been described referring to the example of the enclosed drawing, it will be obvious that the invention is not limited to that but it may be modified in many ways within the scope of the attached claims.

Claims

1. A method for monitoring an object, c h a r a c t e r i z e d in that the method comprises: utilizing a camera system comprising at least one camera disposed in a monitored space to store images of the monitored space, wherein the camera system is suitable for use under any lighting conditions; analyzing (702, 706, 708) images produced by the camera system with signal processing equipment by searching (702) from the stored images at least one object to be monitored and by comparing (706, 708) parameters describing forms and/or motions of the at least one object detected in the images with stored reference parameters corresponding to reference forms and/or reference motions; selecting, on the basis of an event detected in the analysis, an alarm type, wherein each event detectable in the analysis is associated with one of a plurality of different alarm types and each alarm type defines a different alarm protocol comprising instructions as how to carry out the alarm; and conducting (710) an alarm of the selected alarm type as a result of said event and according to the alarm protocol associated with the selected alarm type so as to notify the occurrence of said event.

2. The method of claim 1 , further comprising conducting the alarm by using telecommunication means operably connected to the signal processing equipment.

3. The method of claim 1 or 2, further comprising: defining a priority of an alarm to be conducted from a plurality of priority levels on the basis of the event detected in the analysis, and conducting an alarm according to the defined priority.

4. The method of any preceding claim 1 to 3, wherein the at least one object comprises a human figure, the method further comprising: detecting a request for help by the human figure when comparing the parameters describing postures and/or motions of the human figure with reference parameters of reference postures and/or reference motions associated with the request for help; and conducting a predetermined alarm associated with the request for help as a result of the detection of the request for help.

5. The method of any preceding claim 1 to 4, wherein the at least one object comprises a human figure, the method further comprising: detecting a behavior pattern typical for a determined illness by searching (702) from the stored images a human figure and by comparing (706, 908) parameters describing postures and/or motions of the human figure with stored reference parameters corresponding to reference postures and/or reference motions associated with selected illnesses, conducting (910), on the basis of the behavior pattern detected in the analysis, a predetermined action to notify information on said behavior pattern.

6. The method of claim 5, further comprising: defining a time window within which the frequency of occurrence of detected motions and/or postures associated with the determined illness are monitored; and conducting the predetermined action when the frequency of occurrence exceeds a predetermined threshold within the time window.

7. The method of claim 5 or 6, wherein the predetermined action comprises: registering in database information related to the behavior pattern or conducting an alarm by using a telecommunication unit (200) operationally connected to the signal processing equipment.

8. The method of any preceding claim 5 to 7, wherein the predetermined action comprises: storing in the database information of a posture and/or motion related to the detected determined illness, wherein the stored information illustrates the motion of the human figure for a defined duration starting from a time instant before said detection and ending at a time instant after said detection.

9. The method of any preceding claim 1 to 9, further comprising: constructing (705), by using signal processing equipment, a three- dimensional model of a human figure detected in an image captured with the camera system; analyzing the three-dimensional model of the human figure by comparing (706, 708) parameters describing postures and/or motions of the three-dimensional model modeling the human figure with stored reference parameters corresponding to reference postures and/or reference motions; conducting (710), as a result of an event detected in the analysis, a predetermined action to notify information related to said event.

10. The method of any preceding claim 1 to 10, further comprising: defining a special zone (600) in the stored images; utilizing parameters corresponding to one or more reference forms and/or reference motions associated with the special zone after a human figure is detected entering the special zone; conducting an alarm if the human figure is detected exiting the special zone in a manner deviating from the utilized reference motion and/or reference posture.

1 1 . The method of any preceding claim 1 to 1 1 , further comprising: defining a special zone (600) in the stored images; utilizing parameters corresponding to one or more reference postures and/or reference motions associated with a human figure entering the special zone; conducting an alarm if a human figure is detected entering the special zone in a manner deviating from the utilized reference motion and/or reference posture.

12. The method of any preceding claim 1 to 12, further comprising: acquiring audio signals from at least one microphone disposed in the monitored space; analyzing the acquired audio signals so as to detect determined voice patterns in the acquired audio signals; selecting, on the basis of a sound event detected in the audio analysis, an action to be performed, wherein each event detectable in the audio analysis is associated with one of a plurality of different actions; and conducting (710) the selected action as a result of said sound event.

13. The method of claim 13, wherein an audio signal associated with canceling an alarm is detected in the analysis and wherein the action to be performed is canceling the alarm that has already been made.

14. The method of claim 12 or 13, wherein the analysis further comprises detecting a behavior pattern typical for a determined illness by detecting, with voice recognition equipment, voice patterns in the acquired audio signals and by comparing (706, 908) parameters describing the detected voice patterns with stored reference parameters corresponding to reference voice patterns associated with selected illnesses.

15. A system for monitoring an object, comprising: a camera system (100, 102) suitable to be disposed in a monitored space and arranged to produce images of the monitored space, wherein the camera system is suitable for use under any lighting conditions; signal processing equipment (202) configured to analyze (702, 706, 708) images produced by the camera system by searching (702) from the images at least one object to be monitored and by comparing (706, 708) parameters describing forms and/or motions of the at least one object detected in the images with stored reference parameters corresponding to reference forms and/or reference motions; and a control unit (204) configured to select, on the basis of an event detected in the analysis, an alarm type, wherein each event detectable in the analysis is associated with one of a plurality of different alarm types and each alarm type defines a different alarm protocol comprising instructions as how to carry out the alarm, and to conduct (710) an alarm of the selected alarm type as a result of said event and according to the alarm protocol associated with the selected alarm type so as to notify the occurrence of said event.

16. The system of claim 15, wherein the control unit (204) is further configured to conduct the alarm by using a telecommunication unit (200) operably connected to the control unit (204).

17. The system of claim 15 or 16, wherein the control unit (204) is further configured to define a priority of an alarm to be conducted from a plurality of priority levels on the basis of the event detected in the analysis, and to conduct an alarm according to the defined priority.

18. The system of any preceding claim 15 to 17, wherein the at least one object comprises a human figure, and the signal processing equipment is further configured to detect a request for help by the human figure when comparing the parameters describing postures and/or motions of the human figure with reference parameters of reference postures and/or reference motions associated with the request for help; and the control unit is further configured to conduct a predetermined alarm associated with the request for help as a result of the detected request for help.

19. The system of any preceding claim 15 to 18, wherein the at least one object comprises a human figure, and the signal processing equipment is further configured to detect a behavior pattern typical of a determined illness by searching (702) from the stored images a human figure and by comparing (706, 908) parameters describing postures and/or motions of the human figure with stored reference parameters corresponding to reference postures and/or reference motions associated with selected illnesses; and the control unit is further configured to conduct (910), on the basis of the behavior pattern detected in the analysis, a predetermined action to notify information on said behavior pattern.

20. The system of claim 19, wherein the signal processing equipment is further configured to define a time window within which the frequency of occurrence of detected motions and/or postures associated with the determined illness are monitored; and the control unit is further configured to conduct the predetermined action when the frequency of occurrence exceeds a predetermined threshold within the time window.

21 . The system of claim 19 or 20, wherein the control unit is configured to register as the predetermined action in database information related to the behavior pattern or conducting an alarm by using a telecommunication unit (200) operationally connected to the control unit.

22. The system of any preceding claim 19 to 21 , wherein the control unit is configured to store as the predetermined action in the database information of a posture and/or motion related to the detected determined illness, wherein the stored information illustrates the motion of the human figure for a defined duration starting from a time instant before said detection and ending at a time instant after said detection.

23. The system of any preceding claim 15 to 22, wherein the signal processing equipment is further configured to construct (705) a three- dimensional model of a human figure detected in an image captured with the camera system and to analyze the three-dimensional model of the human figure by comparing (706, 708) parameters describing postures and/or motions of the three-dimensional model modeling the human figure with stored reference parameters corresponding to reference postures and/or reference motions; and the control unit is further configured to conduct (710), as a result of an event detected in the analysis, a predetermined action to notify information related to said event.

24. The system of any preceding claim 15 to 23, wherein the signal processing equipment is further configured to define a special zone (600) in the stored images, to utilize parameters corresponding to one or more reference forms and/or reference motions associated with the special zone after a human figure is detected entering the special zone; and the control unit is further configured to conduct an alarm if the human figure is detected exiting the special zone in a manner deviating from the utilized reference motion and/or reference posture.

25. The system of any preceding claim 15 to 24, wherein the signal processing equipment is further configured to define a special zone (600) in the images and to utilize parameters corresponding to one or more reference postures and/or reference motions associated with a human figure entering the special zone; and the control unit is further configured to conduct an alarm, if a human figure is detected entering the special zone in a manner deviating from the utilized reference motion and/or reference posture.

26. The system of any preceding claim 15 to 25, further comprising at least one microphone disposed in the monitored space and operably connected to the signal processing equipment, wherein the signal processing equipment is further configured to acquire audio signals the at least one microphone and to analyze the audio signals so as to detect determined voice patterns in the acquired audio signals; and the control unit is further configured to select, on the basis of a sound event detected in the audio analysis, an action to be performed, wherein each event detectable in the audio analysis is associated with one of a plurality of different actions, and to conduct (710) the selected action as a result of said sound event.

27. The system of claim 26, wherein the signal processing equipment is configured to detect an audio signal associated with canceling an alarm and wherein the control unit is configured to cancel the alarm that has already been made in response to the detection of the audio signal associated with canceling the alarm.

28. The system of claim 26 or 27, wherein the signal processing equipment is further configured to detect a behavior pattern typical of a determined illness by detecting, with voice recognition equipment, voice patterns in the acquired audio signals and by comparing (706, 908) parameters describing the detected voice patterns with stored reference parameters corresponding to reference voice patterns associated with selected illnesses.

29. A computer program product readable by a computer which, when executed by the computer, configures the computer to execute a computer process comprising: receiving one or more camera images of a monitored object from a camera system comprising at least one camera disposed in a monitored space to store images of the monitored space, wherein the camera system is suitable for use under any lighting conditions; analyzing (702, 706, 708) images produced by the camera system by searching (702) from the stored images at least one object to be monitored and by comparing (706, 708) parameters describing forms and/or motions of the at least one object detected in the images with stored reference parameters corresponding to reference forms and/or reference motions; selecting, on the basis of an event detected in the analysis, an alarm type, wherein each event detectable in the analysis is associated with one of a plurality of different alarm types and each alarm type defines a different alarm protocol comprising instructions as how to carry out the alarm; and conducting (710) an alarm of the selected alarm type as a result of said event and according to the alarm protocol associated with the selected alarm type so as to notify the occurrence of said event.

30. A system comprising means for carrying out the method according to any preceding claim 1 to 14.