JP2008123360A - Device, method, and program for extracting/determining human body specific area - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To efficiently and accurately extract specific areas in human body areas of a plurality of persons. <P>SOLUTION: The device for extracting/determining a human body specific area is provided with human body area detection means 2, 12, and 22 for detecting a human area in a photographed image of a person, human body determination means 3, 13, and 23 for associating a human body area in a current frame detected by the human body area detection means with that detected in the previous frame, and human body specific area extraction means 4, 14, and 24 for extracting a specific area from the human body area determined as a human body area by the human body determination means. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an image processing technique for detecting a specific part of a body corresponding to each person from a real-time image in which one or more persons exist.

Currently, there are various information presentation devices such as large displays and portable terminals placed in public spaces. As one of the techniques, a technique capable of detecting a palm of a person and presenting an image on the palm of the palm has been proposed (see Non-Patent Document 1).
According to this technology, in an information presentation apparatus that can present an image on the palm of a person, correspondence between the person and the hand is performed using only information of one frame of the image captured by the camera.

In this case, considering that there are a plurality of people who are moving at the same time, the human body that existed in one frame is correctly detected as to which human body is associated in the next frame, and the efficiency of the human body region is detected. In particular, it is necessary to extract a specific region of the human body (for example, a hand region or a forehead region).
In addition, when hand regions are extracted from a plurality of human body regions, it is preferable to accurately determine whether the hand region is the left or right hand even if the position of the extracted hand continues to move.
Furthermore, while specific areas (for example, hand areas) of several human bodies move in a complex manner, it is necessary to correctly associate specific areas of the same human body in a certain frame and the next frame.

Yoko Ishii, Satoshi Kobayashi, Akira Nakayama, Michi Hosoda, “Hand Display Interface for Directing Encounters with Information,” IEICE Tech. Bulletin, MVE-2005-42, Sep. 2005

  An object of the present invention is to provide a method for efficiently and correctly extracting a specific region of a human body region of a plurality of people.

  A human body specific region extraction / determination device according to an aspect of the present invention includes a human body region detection unit that detects a human region from a photographed person image, and a human body in a current frame detected by the human body region detection unit. Human body determination means for associating a region with a human body area detected in a frame prior to the present, and human body identification for extracting a specific area from the human body areas determined by the human body determination means And a region extracting means. The present invention is not limited to an apparatus, and can also be realized as an invention of a method or program realized by the apparatus.

  According to the present invention, it is possible to efficiently and correctly extract specific regions of human body regions of a plurality of people.

  Embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing a schematic configuration of a human body specific region extraction / determination device according to an embodiment of the present invention. FIG. 1 also shows the flow of data exchanged in each block. In the present embodiment, an arbitrary area can be designated as the specific area of the human body to be detected. The present embodiment will be described assuming that the specific area is a hand area.

  The human body specific region extraction / determination device according to the present embodiment includes at least a photographing unit 1, a human body region detection unit 2, a human body determination unit 3, a human body specific region extraction unit 4, a human body specific region determination unit 5, A storage unit 6, a storage unit 7, and a left / right determination unit 8 are provided.

  The photographing unit 1 photographs a human body from directly above and transmits a photographed image obtained by photographing to the human body region detecting unit 2. The photographing unit 1 is a camera, and parameters such as the shutter speed and exposure of the camera are fixed.

  The human body region detection unit 2 detects a person region from the photographed image sent from the photographing unit 1. In this case, a plurality of person regions may be detected. Then, the human body region detection unit 2 sends the detected human body region information (this human body region information is referred to as “human body region information of the current frame”) to the human body determination unit 3.

  The storage unit 6 stores the human body region information of the previous frame, and sends the human body region information of the previous frame to the human body determination unit 3. The storage unit 6 receives the human body region information of the current frame output from the human body region detection unit 2 to the human body determination unit 3, and stores the human body region information of the next frame as the human body region information of the previous frame. To do. The previous frame may be the previous frame or a frame several frames before, and these are collectively referred to as “previous frame” in this specification.

  The human body determination unit 3 acquires the human body region information of the current frame sent from the human body region detection unit 2 and the human body region information of the previous frame sent from the storage unit 6. The human body determination unit 3 obtains the centroids of the previous frame and the current frame, and compares the centroid distance between the human body region existing in the current frame and the human body region of the previous frame when the human body exists in the previous frame. Note that the method of obtaining the center of gravity of the previous frame and the current frame may be, for example, using the geometric center of the detected human body region as the center of gravity. When the center of gravity of the human body region of the current frame and the center of gravity of the human body region of the previous frame are closer than a predetermined distance (that is, below a predetermined battle value), the human body determination unit 3 And the corresponding human body region in the previous frame are determined as the same human body region. Further, when the human body region of the current frame does not correspond to the human body region of the previous frame, the human body determination unit 3 determines that the human body region is a newly appearing human body region. The human body determination unit 3 performs the same processing as described above for the number of human body regions existing in the current frame. The human body determination unit 3 sends the associated human body region information to the human body specific region extraction unit 4 and the human body specific region determination unit 5.

  The human body specific region extraction unit 4 extracts a hand region from the human body region information sent from the human body determination unit 3. When there are a plurality of human body regions sent from the human body determination unit 3, the human body specific region extraction unit 4 extracts a hand region (hand region) from each human body region information of the plurality of human body region information. Since the hand region is at the end of the human body region, it is preferable to extract a maximum of two regions at the end of the body for each human body region. Then, the human body specific region extraction unit 4 transmits the extracted hand region information to the left / right determination unit 8 and the human body specific region determination unit 5.

  The left / right determination unit 8 uses the information sent from the human body specific region extraction unit 4 to determine which hand is the left or right hand. The left / right determination unit 8 transmits left / right information of the determined hand region to the human body specific region determination unit 5.

  The storage unit 7 stores information on the hand region of the previous frame and sends the information on the hand region of the previous frame to the human body specific region determination unit 5. Further, the storage unit 7 receives the information of the hand region of the current frame output from the human body specific region extraction unit 4 to the human body specific region determination unit 5, and uses the hand region of the next frame as the information of the hand region of the previous frame. Remember until you receive the information.

The human body specific region determination unit 5 includes the current frame hand region information transmitted from the human body specific region extraction unit 4, the current frame left / right determination information transmitted from the left / right determination unit 8, and the previous information transmitted from the storage unit 7. Get hand region information of the frame. Then, the human body specific region determination unit 5 performs the following processing for each human body region information existing in the current frame, and outputs the processing result.
(1) When a hand region exists in the previous frame and a hand region exists in the current frame, the angle formed between the straight line connecting the center of gravity of each hand region and the center of gravity of the human body and the reference straight line is compared. . When the difference in angle between the previous frame and the current frame is within a predetermined threshold, the corresponding hand region of the previous frame and the hand region of the current frame are determined as the same hand region.
(2) If the hand area of the current frame does not correspond to the hand area of the previous frame, the hand area is determined as a newly appearing hand area.
As described above, the hand region as the specific region of the human body is determined.

  According to the present embodiment, it is possible to reduce misrecognition by using the human body region information of the past frame (previous frame), and to correctly associate the human body regions of a plurality of persons who move with each other in time series. Furthermore, in the process of detecting a specific area (for example, “hand area”) from human body area information, it is possible to detect the specific area more efficiently by reducing erroneous recognition that detects areas other than the specific area. become.

  Hereinafter, as a specific example, a case where a hand region is detected as a specific region (first specific example) and a case where a forehead region is detected as a specific region (second specific example) will be described.

(First specific example)
FIG. 2 is a diagram illustrating a configuration example of the human body specific region extraction / determination device according to the first specific example. In this specific example, the specific area detected from the human body area information will be described as a hand area.
In the first specific example, the photographing unit 1 in the above embodiment is implemented as a video camera 10. Further, the computer 11 includes a human body region detection unit 12, a human body determination unit 13, a human body specific region extraction unit 14, a left / right determination unit 17, a human body specific region determination unit 15, and a storage unit 16. Yes. The video camera 10 is installed above the user's head and acquires the user's image from directly above the user. The video camera 10 is connected to the computer 11 and transmits the captured image to the computer 11.

  The image sent to the computer 11 is input to the human body region detection means 12. The human body region detection means 12 receives an image taken by the video camera 10. First, the human body region detection unit 12 obtains a difference between an image when a person is not present in the shooting region of the video camera 10 and a current image (that is, an image of the current frame). Then, when a region having a size equal to or larger than a predetermined threshold exists in the image obtained by the difference, the human body region detection unit 12 detects this region as a human body region. When the human body region is detected, the human body region detection unit 12 obtains the size and position of the center of gravity of the detected region, the size and position of the bounding box, and sends the information to the human body determination unit 13 as human body region information. Note that the bounding box may have any shape as long as it is a polyhedron surrounding the detected area, and can be obtained by a known technique, and thus the description thereof is omitted.

  The human body determination means 13 determines whether or not the extracted human body region is a human body. A specific processing flow of the human body determination means 13 is shown in FIG. FIG. 3 is a flowchart showing a specific processing flow of the human body determination means 13.

  The human body determination unit 13 acquires information on the human body region existing in the current frame from the human body region detection unit 12. Then, the human body determination unit 13 determines whether a human body region is detected in the current frame (step A1). In step A1, if no human body area is detected in the current frame, the process proceeds to step A10. In step A1, if a human body region is detected in the current frame, the processing from step A3 to step A8 is repeated for the number of human body regions detected in the current frame. Step A2 indicates the start of the loop, and step A9 indicates the end of the loop.

  Specifically, first, when the human body is not detected in the previous frame (No in step A3), the human body determination unit 13 determines that the human body in the current frame is a newly appearing human body region. Then, the processing from step A3 to step A8 is performed for the next human body region.

  First, the human body determination unit 13 obtains the center-of-gravity position information of the human body region existing in the previous frame from the storage unit 16, and when the human body region is detected in the previous frame (Yes in step A3), It is determined whether there is a human body region that is a candidate for association among the human body regions in the previous frame (step A4). If there is a human body region that is a candidate for association among the human body regions in the previous frame (Yes in step A4), the centroid positions of all human body regions in the previous frame that can be associated and the centroid positions of one human body region in the current frame Is measured (step A5). If the determination result in step A3 and step A4 is No, the process proceeds to step A9.

  The human body determination means 13 determines whether or not one human body region of the previous frame having the shortest distance among the measured distances of the center of gravity positions is equal to or less than a predetermined battle value (step A6). If it is equal to or less than the threshold value, one human body region of the previous frame is associated with one human body region of the current frame (step A7). If there are a plurality of human body regions in the current frame, one human body region in the previous frame that has been associated is excluded from the processing target (step A8). Then, the processing from step A3 to step A8 is repeated, and the association with the human body region of the remaining current frame is determined among the remaining previous frame regions. As a result, the associated human body regions have a one-to-one relationship between the human body region of the previous frame and the human body region of the current frame. If there is no more human body area in the previous frame to be associated, it is determined that the remaining human body area in the current frame is a newly appearing human body area.

  The human body determination unit 13 transmits information such as the size and position of the center of gravity of the human body region that has been determined to be associated, and the size and position of the bounding box to the human body specific region extraction unit 14 and the human body specific region determination unit 15. The human body determination means 13 includes the size and position of the center of gravity of the human body region for which the determination of the association is made, the size and position of the bounding box, which region of the previous frame has been associated (or has not been performed), etc. Information is transmitted to the storage means 16. The storage means 16 updates the data in the case of a human body region that has been associated, stores the data as a human body region that has not been associated, and is associated with the current frame. The data of the previous frame that did not exist is deleted (step A10).

  As a subsequent process (step A11), the human body specific region extracting unit 14 receives information on the human body region for which the association is determined from the human body determining unit 13. Then, the human body specific region extracting means 14 extracts a skin color region from the received information on the human body region. As a method of extracting the skin color region, HSV conversion, which is one of the general processes that can decompose an image into a certain component, is performed, and the H component (color component) and the S component (saturation component) are respectively determined from certain threshold values. Means for extracting a region smaller than a large threshold value may be used. The human body specific region extracting means 14 extracts a region that is larger than a certain battle value and smaller than the battle value from the extracted skin color region, and uses this as a hand candidate region. The human body specific region extracting unit 14 obtains the center of gravity of the hand candidate region. The human body specific area extracting means 14 obtains the distance between each hand candidate area and the center of gravity of the human body, and extracts a maximum of two hand candidate areas having the center of gravity farthest from the human body center of gravity as the hand areas. The human body specific region extracting unit 14 sends the size of the detected hand region, the position / size of the bounding box, the gravity center position information, and the like to the left / right determining unit 17.

The left / right determination means 17 receives the hand area information sent from the human body specific area extraction means 14 and performs the following processing.
If no hand is present in the previous frame and only one hand is present in the current frame, that hand is determined as the right hand. If no hand is present in the previous frame and both hands are present in the current frame, a straight line connecting the hand and the center of gravity of the human body region is obtained, and the hand on the right side of the acute angle (θ in FIG. 4) formed by the straight line is obtained. As the right hand and the left hand as the left hand (see FIG. 4).

  If a hand is present in the previous frame and only one hand is present in the current frame, that hand is determined as the right hand. If a hand is present in the previous frame and both hands are present in the current frame, a straight line connecting the hand and the center of gravity of the human body region is obtained, and the right hand of the right side of the acute angle (θ in FIG. 4) formed by these straight lines is defined as the right hand. The left hand is determined as the left hand (see FIG. 4).

  The left / right determination unit 17 transmits left / right determination information to the human body specific region determination unit 15. The human body specific region determination unit 15 associates the human body specific region with the human body region according to the flowchart shown in FIG. FIG. 5 is a flowchart showing the flow of processing of the human body specific region determination unit 15. With reference to FIG. 5, the flow of processing of the human body specific region determination unit 15 will be described.

The human body specific region determination unit 15 receives information on the human body region associated with the previous frame and the current frame from the human body determination unit 13. The human body specific region determination unit 15 receives the left / right determination information from the left / right determination unit 17. The human body specific area determination unit 15 acquires from the storage unit 16 the center-of-gravity position information and left-right information of the hand area that existed in the previous frame, and information on which human body area is the hand area. Then, the processing from step B2 to step B17 is repeated for each human body region. Step B1 indicates the start of the loop, and step B18 indicates the end of the loop.
The human body specific region determination means 15 determines whether or not the human body region in which the hand region of the previous frame is detected is associated (step B2), and according to the result, the human body region is divided into the following conditions, To perform the corresponding process.

1) When there is no hand area in the current frame (No in step B3, No in step B5)
If no hand area exists even in the previous frame (No in step B3), the process of step B18 is advanced. If a hand has been detected in the previous frame (No in step B5), the hand is recognized as a lost hand, and the hand area of the frame is deleted (step B6). It is assumed that there is no other hand.

2) When a hand is present in the current frame and a hand is not detected in the previous frame (Yse in step B3)
The hand of the current frame is recognized as a newly appearing hand, and the left and right information of the hand is used as it is (step B4).

3) When both hands are detected in the previous frame and both hands are detected in the current frame (step B7)
Using the right hand and left hand information of the current frame as they are, the right hand existing in the previous frame is associated with the right hand of the current frame, and the left hand existing in the previous frame is associated with the left hand of the current frame (step B8).

4) When one hand is detected in the previous frame and both hands are detected in the current frame (step B9)
Based on FIG. 6, the angle a and the angle b of the right and left hands of the current frame are compared with the angle a (angle b) of the hand of the previous frame. , It is determined that the hand is present from the previous frame, and the center-of-gravity coordinates, angle, and left / right information of the hand region are updated (step B10). At this time, priority is given to the left and right information of the current frame even if the left and right information of the hand of the current frame is different from the left and right information of the hand of the previous frame.
Note that one hand of the current frame that has not been associated is recognized as a newly appearing hand (step B11).
In FIG. 6, a line passing through the center of gravity of the human body region and parallel to the horizontal axis direction of the camera image is defined as a reference line C. An angle formed by the reference line C in FIG. 6 and “a line connecting the center of gravity of the human body region with the center of gravity of the right hand” is an angle a. An angle formed by the reference line C in FIG. 6 and “a line connecting the centroid point of the human body region with the centroid point of the left hand” is an angle b.

5) When both hands are detected in the previous frame and one hand is detected in the current frame (step B12)
If only one hand is detected in the current frame, that hand is determined as the right hand. Compare the right hand angle a of the current frame with the right hand and left hand angles a and b of the previous frame, associate the hand of the front frame with the closer angle, and match the center of gravity coordinates and angle of the hand region Information is updated (step B13). At this time, if the left and right information is different between the previous frame and the current frame, priority is given to the left and right information of the previous frame. One hand of the previous frame that has not been associated is recognized as a hand that has disappeared from the screen (step B14).

6) When one hand is detected in the previous frame and one hand is detected in the current frame (step B15)
Based on FIG. 6, if the hand angle of the current frame is compared with the hand angle of the previous frame, and if it is less than the battle value, the current frame and the hand of the previous frame are associated, The angle and left / right information is updated (step B16). The left and right information uses information on the hand of the previous frame. If the angle difference is larger than the battle value, the hand of the previous frame is recognized as a hand that has disappeared from the screen, and the hand of the current frame is recognized as a newly appearing hand (step B17). As for the left / right information of the hand, information opposite to the left / right information of the previous frame is associated with the hand of the current frame.

  The human body specific region determination means 15 performs the association with the size and position of the specific region after the determination, the size and position of the bounding box, left and right information, information on which human body region is the hand region, and the region of the previous frame. Information such as whether or not it has been sent is transmitted to the storage means 16. Thereby, the hand determination process ends (step B19).

  The storage means 16 updates the data in the case of the human body specific area in which the association is performed, stores the data as a human body specific area that has newly appeared in the case of the human body specific area in which the association has not been performed, and specifies the current frame Data in the previous frame specific area that has not been associated with the area is deleted.

(Second specific example)
FIG. 7 is a diagram illustrating a configuration example of the human body specific region extraction / determination device according to the second specific example. In this specific example, the specific area detected from the human body area information is assumed to be a forehead area.
In the second specific example, the photographing unit 1 in the above embodiment is implemented as a video camera 20. The computer 21 includes a human body region detection unit 22, a human body determination unit 23, a human body specific region extraction unit 24, and a storage unit 26. The video camera 20 is installed above the user's head and acquires the user's image from directly above the user. Further, the video camera 20 is connected to a computer 21 and transmits a captured image to the computer 21.

  The image sent to the computer 21 is input to the human body region detection means 22. The human body region detection means 22 receives an image taken by the video camera 20. First, the human body region detection unit 22 calculates a difference between an image when a person is not present in the shooting region of the video camera 20 and a current image (that is, an image of the current frame). Then, when a region having a size equal to or larger than a predetermined threshold exists in the image obtained by the difference, the human body region detection unit 22 detects this region as a human body region. When the human body region is detected, the human body region detection unit 22 obtains the size and position of the center of gravity of the detected region, the size and position of the bounding box, and sends the information to the human body determination unit 23 as human body region information.

The human body determination unit 23 determines whether or not the extracted human body region is a human body. Since the specific processing of the human body determination unit 23 is the same as that of the human body determination unit 13 of the first specific example, description thereof is omitted.
The human body determination unit 23 transmits the information on the human body region for which the association is determined to the human body specific region extraction unit 24 and the storage unit 26.

  The human body specific region extracting unit 24 receives information on the human body region for which the association is determined from the human body determining unit 23. The human body specific region extraction unit 24 extracts a skin color region from the received information on the human body region. As a method for extracting a skin color region, as in the first specific example, HSV conversion, which is one of the general processes that can decompose an image into certain components, is performed, and an H component (color component) and an S component (saturation) are performed. It is also possible to use means for extracting a region where the degree component) is larger than a certain threshold value and smaller than a certain threshold value. The human body specific region extracting means 24 extracts a region that is larger than a certain battle value and smaller than the battle value from the extracted skin color region, and sets this as a forehead candidate region.

Then, the human body specific region extraction unit 24 obtains the center of gravity point of the forehead candidate region. The human body specific area extracting means 24 obtains the distance between each forehead candidate area and the center of gravity of the human body, and among the distances closer than a certain threshold value, “one forehead candidate area having the center of gravity closest to the body center of gravity” Is extracted as the forehead area. When the distance between each forehead candidate area and the center of gravity of the human body does not satisfy a certain threshold value, the human body specific area extraction unit 24 determines that the forehead candidate area is not a forehead. When extracting the forehead candidate area as the forehead area, the distance from the center of gravity of the human body is examined because the forehead area exists near the center of gravity of the human body when shooting the human body from the camera installed on the user's head, This is because the forehead region can be correctly extracted by measuring the distance between the forehead candidate region and the center of gravity of the human body.
The human body specific area extraction unit 24 transmits the information of the extracted forehead area to the storage unit 26.

As described above, in the present embodiment, information on the human body area in the past frame (that is, the previous frame) is stored, and the number of human body specific areas for each detected human body area and the specific area are determined. Using the location information when it exists in the human body, the region that meets those conditions is extracted. For example, when the specific region is a hand region, since there are a maximum of two palms in the human body and at the end of the body, a maximum of two regions at the end of the body are extracted for each human body region. As another example, when the specific area is the forehead, there is at most one in the human body, and when the human body is photographed from directly above, it is located near the center of gravity of the body, A maximum of one region near the center of gravity is extracted for each human body region.
In the left / right determination of the hand region, the left / right determination is performed by comparing the angles formed by the two straight lines connecting the center of gravity of the human body region and the center of gravity of the hand region of each of the right hand and the left hand with a certain reference line. As a result, it is possible to correctly determine the right hand and the left hand even if the hand regions of a plurality of people move in a complicated manner.
In addition, the angle information connecting the center of gravity of the specific area of the human body and the center of gravity of the human body area in a certain frame is stored, and by using this information, the specific areas of the human bodies of the plurality of persons to be moved are correctly associated. . As a result, even if the specific areas of the human bodies of a plurality of people move in a complex manner, it is possible to correctly associate the specific areas of the human body between a certain frame and the next frame.

The present invention is not limited to the above-described embodiments, and various modifications can be made without departing from the scope of the invention at the stage of implementation. Further, the above embodiments include inventions at various stages, and various inventions can be extracted by appropriately combining a plurality of disclosed constituent elements.
In addition, for example, even if some structural requirements are deleted from all the structural requirements shown in each embodiment, the problem described in the column of the problem to be solved by the invention can be solved, and the effect described in the effect of the invention Can be obtained as an invention.

It is a block diagram which shows schematic structure of the human body specific area | region extraction and determination apparatus which concerns on one Embodiment of this invention. It is a figure which shows the structural example of the human body specific area | region extraction and determination apparatus which concerns on a 1st specific example. 5 is a flowchart showing a specific processing flow of human body determination means 13; The figure for demonstrating an example of the principle which determines the right and left of a hand area. 4 is a flowchart showing a flow of processing of a human body specific area determination unit 15. The figure for demonstrating an example of the principle which determines matching with the front frame of the right and left of a hand area, and the present frame. It is a figure which shows the structural example of the human body specific area | region extraction and determination apparatus which concerns on a 2nd specific example.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Imaging | photography part 2 ... Human body area | region detection part 3 ... Human body determination part 4 ... Human body specific area extraction part 5 ... Human body specific area determination part 6 ... Memory | storage part 7 ... Memory | storage part 8 ... Left-right determination part 10 ... Video camera 20 ... Video camera DESCRIPTION OF SYMBOLS 11, 21 ... Computer 12, 22 ... Human body area | region detection means 13, 23 ... Human body determination means 14, 24 ... Human body specific area extraction means 15 ... Human body specific area determination means 16, 26 ... Storage means 17 ... Left-right determination means

Claims (8)

Human body region detecting means for detecting a person region from a photographed person image;
Human body determination means for associating the human body area in the current frame detected by the human body area detection means with the human body area detected in a frame before the current time;
A human body specific region extraction / determination device comprising: a human body specific region extraction unit that extracts a specific region from the human body region determined to be a human body region by the human body determination unit. 2. The human body specific region extracting / determining device according to claim 1, wherein the human body determining means determines whether the detected human body region in the current frame is present from a previous frame. Specific area extraction / determination device. 3. The human body specific region extraction / determination device according to claim 1 or 2, further comprising storage means for storing the information obtained by the human body determination means until processing of the next frame. Specific area extraction / determination device. The human body specific region extraction / determination device according to any one of claims 1 to 3, wherein the detected hand region when the specific region extracted by the human body specific region extraction unit is a hand region. A human body specific region extracting / determining device further comprising left / right determining means for determining whether the hand is a left or right hand by an angle of a straight line connecting the center of gravity of the human body and the center of gravity of the hand region. 5. The human body specific region extraction / determination device according to claim 1, wherein when there are a plurality of the specific regions extracted by the human body specific region extraction unit, the specific region of the current frame is determined. A human body specific area extracting / determining device further comprising human body specific area determining means for performing association determination based on a center-of-gravity distance between a current area and a specific area of a previous frame. 6. The human body specific region extracting / determining apparatus according to claim 1, further comprising a photographing unit for photographing a person image from directly above. apparatus. Detect the person's area from the photographed person's image,
The human body region in the current frame detected by the human body region detection means is associated with the human body region detected in a frame before the current time,
A human body specific area extraction / determination method characterized by extracting a specific area from the associated human body areas. Detect the person's area from the photographed person's image,
The human body region in the current frame detected by the human body region detection means is associated with the human body region detected in a frame before the current time,
A computer-readable human body specific region extraction / determination program characterized by extracting a specific region from the associated human body regions.

Images