JP2018078528A - Image analysis device, image analysis method, and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image analysis device, an image analysis method, and a program capable of easily obtaining a clear image when imaging is performed close to a human body surface.SOLUTION: An image analysis device includes an imaging unit that captures image data of two or more frames at a predetermined frame rate, a control unit that fixes a focal distance of the imaging unit by disabling an autofocus function of the imaging unit, starts imaging in a state in which the imaging unit is located in a first distance different from the focal distance for the imaged portion of a human body surface, and continues the imaging until the imaging unit moves to a location at which the imaging unit exceeds the focal distance with respect to the imaged portion and reaches the second distance, an obtaining unit that obtains image data in which the energy of a frequency component based on the sharpness of the imaged portion is maximum from the image data of two or more frames, and an analysis unit that analyzes the image data that has been obtained by the obtaining unit.SELECTED DRAWING: Figure 3

Description

  The present invention relates to an image analysis apparatus, an image analysis method, and a program, and more particularly to an image analysis apparatus, an image analysis method, and a program for measuring a human body surface.

  A technique is known in which a human body surface is imaged using a camera provided in a smartphone, a tablet terminal, or the like, and the state of the human body surface is analyzed based on the obtained image. Patent Document 1 discloses an application program that captures an image of a face using a smartphone camera and detects skin spots or wrinkles from the face image. Patent Document 2 discloses an application program that captures skin using a camera of a tablet terminal and calculates the amount of melanin deposition based on the luminance of the skin image.

JP 2014-219781 A Japanese Patent Laying-Open No. 2015-173762

  In the techniques disclosed in Patent Documents 1 and 2, it is necessary to perform imaging by bringing the camera close to the human body surface. A camera usually has an autofocus function and can easily capture a clear image. However, when the distance between the human body surface and the camera is very short, the autofocus function does not operate properly, and it may be difficult to capture a clear image.

  The present invention has been made in view of such a problem, and provides an image analysis apparatus, an image analysis method, and a program capable of easily acquiring a clear image when imaging close to a human body surface. The purpose is to do.

  An image analysis apparatus according to the present invention fixes an in-focus distance of an image capturing unit that captures image data of a plurality of frames at a predetermined frame rate, and invalidating an autofocus function of the image capturing unit. The imaging unit starts the imaging in a state where the imaging unit is located at a first distance different from the in-focus distance with respect to the imaging site on the human body surface, and the imaging unit sets the in-focus distance to the imaging site. The control unit that continues the imaging until it moves to a position that becomes the second distance beyond, and the image data with the maximum frequency component energy based on the sharpness of the imaging region is acquired from the image data of the plurality of frames And an analysis unit that analyzes the image data acquired by the acquisition unit.

  The image analysis method according to the present invention includes a step of fixing a focusing distance of the imaging unit by disabling an autofocus function of the imaging unit, and the focusing distance of the imaging unit with respect to an imaging site on a human body surface A step of starting imaging of image data of a plurality of frames at a predetermined frame rate in a state where the imaging unit is positioned at a first distance different from the first distance; The step of continuing the imaging until moving to the position of the distance, obtaining the image data having the maximum frequency component energy based on the sharpness of the imaging region from the image data of the plurality of frames, Analyzing the acquired image data.

  According to the present invention, it is possible to easily acquire a clear image when performing imaging in the vicinity of the human body surface, and it is possible to perform accurate analysis using the clear image.

1 is an external view of an image analysis apparatus according to a first embodiment. 1 is a block diagram of an image analysis apparatus according to a first embodiment. It is a functional block diagram of the measurement application program which concerns on 1st Embodiment. 3 is a flowchart of an image analysis method according to the first embodiment. It is for demonstrating the spatial frequency image which concerns on 1st Embodiment. It is a flowchart which shows the detail of the imaging process which concerns on 1st Embodiment. It is a graph which shows the change of the energy of the high frequency component which concerns on 1st Embodiment. It is a figure for demonstrating the extraction method of the high frequency component which concerns on 2nd Embodiment. It is a flowchart of the image analysis method which concerns on 2nd Embodiment.

[First Embodiment]
FIG. 1 is an external view of an image analysis apparatus 100 according to the present embodiment. The image analysis apparatus 100 may be a portable information terminal such as a smartphone, a tablet computer, a PDA (Personal Digital Assistant), a laptop computer, or a mobile phone. Hereinafter, a smartphone will be described as an example of the image analysis apparatus 100. A touch panel 101 is provided on the front surface of the image analysis apparatus 100, and an imaging unit (camera) 102 is provided on the upper back surface. The imaging unit 102 may be provided not only on the back surface of the image analysis apparatus 100 but also on the front surface. The image analysis apparatus 100 is preinstalled with an application program (hereinafter, a measurement application) for measuring the skin condition. The measurement application may be downloaded from a network (not shown) or supplied to the image analysis apparatus 100 through a recording medium such as a memory card.

  An icon of the measurement application is displayed on the touch panel 101, and the measurement application is activated when the user touches the icon. When the user grasps the image analysis apparatus 100 and points the imaging unit 102 toward the subject such as the cheek or the back of the hand, the image analysis apparatus 100 automatically captures the focused image. . The image analysis apparatus 100 can analyze the captured image and display the analysis result on the touch panel 101.

  FIG. 2 is a block diagram of the image analysis apparatus 100 according to the present embodiment. The image analysis apparatus 100 includes an imaging unit 102, a signal processing circuit 204, a timing generation circuit 205, a lens driving circuit 206, an LED 207, a speaker 208, a touch sensor 209, and a coordinate detection circuit 210. The image analysis apparatus 100 further includes a display 211, a display controller 212, a frame memory 213, a CPU 214, a RAM 215, a ROM 216, a storage 217, a motion sensor 218, and a wireless communication module 219.

  The imaging unit 102 includes an optical system 201, an imaging element 202, and an A / D converter 203. The optical system 201 includes an optical filter, a fixed lens, and a focus lens, and forms an image of a subject by forming light from a subject (imaging part) on the imaging surface of the imaging element 202. The imaging device 202 is, for example, a complementary metal oxide semiconductor (CMOS) image sensor or a charge coupled device (CCD) image sensor, and includes a plurality of pixels, color filters, and microlenses arranged two-dimensionally. The plurality of pixels may include imaging pixels and focus detection pixels. The image sensor 202 has an electronic shutter function for controlling the charge accumulation time. Each of the plurality of pixels outputs a pixel signal based on incident light from the optical system 201.

  An A / D (Analog / Digital) converter 203 includes a comparison circuit, a latch circuit, and the like, and converts an analog pixel signal from the image sensor 202 into digital image data. The A / D converter 203 may be provided in the image sensor 202. The timing generation circuit 205 generates a drive signal including a horizontal synchronization signal and a vertical synchronization signal, and outputs the drive signal to the image sensor 202 and the A / D converter 203. The imaging unit 102 can output a moving image having a predetermined frame rate in addition to a still image. The frame rate can be any value such as 1/4 second, 1/30 second, 1/60 second, and the like.

  The signal processing circuit 204 includes a numerical operation circuit and performs digital processing such as white balance adjustment, gamma correction, pixel interpolation, edge enhancement, gradation conversion, noise reduction, and compression on the image data from the A / D converter 203. Perform signal processing. The lens driving circuit 206 includes an actuator or the like, drives the focus lens of the optical system 201, and adjusts the focusing distance. The timing generation circuit 205 outputs timing signals such as a clock signal and a synchronization signal to the image sensor 202 and the A / D converter 203.

  An LED (Light Emitting Diode) 207 is a light source provided in the vicinity of the optical system 201. The LED 207 is used as an illumination unit that irradiates light toward the imaging region in order to obtain illuminance suitable for imaging. The LED 207 may be composed of a plurality of LEDs having different light colors. By adjusting the light emission intensity of each LED, the light having the optimum color temperature and saturation can be emitted to the imaging region. The speaker 208 includes a piezoelectric vibration unit and is driven by a current drive circuit. The speaker 208 outputs a sound signal such as music, a message, and a sound effect, and is used as a notification unit for notifying the user that imaging has been completed.

  The touch sensor 209 is a capacitive sensor having transparent matrix electrodes, and is provided on the display 211. When the user's finger touches the touch sensor 209, the capacitance of the electrode changes. The coordinate detection circuit 210 can detect a change in capacitance in the touch sensor 209 and calculate a position where the user's finger is in contact. The touch sensor 209 is used as an operation unit that receives an instruction from the user.

  The display 211 is, for example, a TFT (Thin Film Transistor) liquid crystal display or an organic EL (Electro Luminescence) display, and displays images, moving images, texts, icons, and the like in accordance with display signals from the display controller 212. The display controller 212 is a processor including a video memory, and controls display on the display 211. The display controller 212 temporarily stores display data from the CPU 214, generates a display signal, and outputs the display signal to the display 211. The touch sensor 20 and the display 211 are integrally formed to constitute the touch panel 101.

  The frame memory 213 can temporarily hold image data of a plurality of frames, and can be used in image processing by the signal processing circuit 204 and the CPU 214. For example, the signal processing circuit 204 may perform camera shake correction and noise reduction by detecting a motion vector in a plurality of frames of image data. In moving image capturing, the frame memory 213 can store image data of a plurality of temporally continuous frames. A part of the RAM 215 may be used as the frame memory 213. A CPU (Central Processing Unit) 214 includes a CPU core, a cache memory, and the like, and comprehensively controls each device of the image analysis apparatus 100. The CPU 214 reads out and executes a predetermined program from the ROM 216 and the storage 217, thereby realizing functions of each unit of the image analysis apparatus 100.

  A RAM (Random Access Memory) 215 is, for example, a DRAM (Dynamic RAM), and is used as a work area of the CPU 214, a program load area, and the like. A ROM (Read Only Memory) 216 is, for example, an EEPROM (Electrically Erasable Programmable ROM), and stores a basic input output system (BIOS), various setting files, and the like. The storage 217 is, for example, a flash memory, and stores basic programs such as an OS (Operating System) and various application programs such as a measurement application. The storage 217 also stores various data such as measurement results by the measurement application, images by the imaging unit 102, and moving images.

  The motion sensor 218 includes an acceleration sensor and a gyro sensor, and detects the movement of the image analysis apparatus 100. The acceleration sensor is composed of a capacitance detection element and the like, and detects acceleration applied to the image analysis apparatus 100. The gyro sensor is composed of a piezoelectric vibration element or the like, and detects the orientation of the image analysis apparatus 100. The image analysis apparatus 100 may further include a GPS (Global Positioning System) sensor, an illuminance sensor, a proximity sensor, and the like.

  The wireless communication module 219 is an interface for performing wireless communication with a communication network such as the Internet. The wireless communication module 219 includes an antenna, and connects to a communication network using a mobile communication system such as LTE (Long Term Evolution), 4G (4th Generation), or a wireless communication system such as a wireless LAN (Local Area Network). The wireless communication module 219 can transmit / receive data to / from an external device via a communication network.

  FIG. 3 is a functional block diagram of the measurement application program according to the present embodiment. The measurement application 300 includes a control unit 301, an acquisition unit 302, and an analysis unit 303. The function of the measurement application 300 is realized by the CPU 214 reading and executing the measurement application 300 stored in the storage 217. That is, the CPU 214 functions as the control unit 301, the acquisition unit 302, and the analysis unit 303 when the measurement application 300 executes the image analysis method.

  The control unit 301 controls operations of the imaging unit (camera) 102, the LED 207, and the speaker 208. The control unit 301 has an autofocus (AF) function and adjusts the focusing distance of the imaging unit 102. In the AF processing, the control unit 301 controls the position of the focus lens of the optical system 201 via the lens driving circuit 206 so that the contrast of the captured image is maximized. The controller 301 is not limited to the AF process using the contrast method, and may perform the AF process using the imaging surface phase difference method. The control unit 301 has an automatic exposure (AE) function, and controls the ISO sensitivity of the image sensor 202 and the speed of the electronic shutter. The control unit 301 can set the AF function and the AE function to be valid or invalid. The control unit 301 further controls the LED 207 and the speaker 208 and acquires an instruction from the user via the touch sensor 209.

  The acquisition unit 302 reads a plurality of frames of image data from the frame memory 213 and acquires the clearest image data from the plurality of frames of image data. The acquisition unit 302 can extract a frequency component based on the sharpness of the imaging region from the image data, and determine the focus of the image data based on the frequency component. The acquisition unit 302 extracts frequency components using frequency transformation such as Fourier transformation. The acquisition unit 302 stores the acquired image data in the storage 217. Data such as a frequency spectrum obtained as a result of the frequency conversion is held in the RAM 215.

  The analysis unit 303 analyzes the image data acquired by the acquisition unit 302 and generates a measurement result. The analysis unit 303 extracts the feature amount of the image data using frequency conversion, and derives a measurement result from the feature amount of the image data. The analysis unit 303 can use the result of frequency conversion by the acquisition unit 302. That is, the analysis unit 303 can extract the feature amount of the image data using the frequency conversion result in the focus determination. As a result, the processing load on the CPU 214 can be reduced and the processing speed can be increased. The analysis unit 303 stores the measurement result in the storage 217. The frequency transform is not limited to Fourier transform, but may be wavelet transform, Laplacian calculation, discrete cosine transform, Hadamard transform, KL (Karhunen-Loeve) transform, or the like.

  FIG. 4 is a flowchart of the image analysis method according to the present embodiment. Here, an example in which the user measures the cheek state using the measurement application 300 will be described. In the image analysis apparatus 100, when the user activates the measurement application 300, the control unit 301 disables the AF function of the imaging unit 102 and fixes the focusing distance to a predetermined distance (step S401). That is, the control unit 301 puts the optical system 201 into a fixed focus state. Here, the in-focus distance can be determined so that the size of the captured image is optimal in the image analysis. For example, the focusing distance may be in the range of 50 mm to 100 mm, or may be the shortest distance that can be set in the imaging unit 102. The control unit 301 sets the light amount, color temperature, and saturation of the LED 207 so that a light environment suitable for imaging is obtained, and turns on the LED 207.

  Subsequently, the control unit 301 starts imaging in a state where the imaging unit 102 is located at the first distance with respect to the user's cheek (step S402). The first distance is a distance different from the focusing distance, and is shorter than the focusing distance, for example. The control unit 301 displays a message such as “Please bring the camera close to the cheek and tap the screen” or “Tap the screen to start shooting. Please move the camera slowly away from the cheek”. The control unit 301 may display an animation together with a message, or may emit a message by voice. When the user brings the imaging unit 102 close to the cheek according to the message and taps the touch panel 101 (see FIG. 1), the control unit 301 starts imaging by the imaging unit 102. Imaging is performed at a predetermined frame rate (for example, 4 to 5 frames per second). Since the distance between the imaging unit 102 and the imaging region at the start of imaging, that is, the first distance is shorter than the in-focus distance, the sharpness of the first image to be captured is low.

  Subsequently, the control unit 301 continues the imaging until the imaging unit 102 moves to a position that is the second distance with respect to the user's cheek (step S403). Here, the second distance is a distance exceeding the focusing distance, and is longer than the focusing distance, for example. After tapping the touch panel 101, the user gradually moves the imaging unit 102 away in a direction perpendicular to the cheek. Based on the change in the sharpness of the image, the control unit 301 determines whether the imaging unit 102 has moved to a second distance that exceeds the in-focus distance with respect to the user's cheek. Note that the control unit 301 may consider that the imaging unit 102 has moved to a position that becomes the second distance when a predetermined time (for example, 6 to 10 seconds) has elapsed.

  Next, the acquisition unit 302 acquires image data having the maximum energy of the high frequency component from the image data of a plurality of frames obtained by imaging (step S404). The image data acquisition process (step S404) may be performed simultaneously with the above-described imaging process (step S403). The acquisition unit 302 decomposes the image data of each frame into frequency components using Fourier transform, and generates a spatial frequency image 500.

  FIG. 5 is a diagram for explaining the spatial frequency image 500. In the spatial frequency image 500, the X axis represents the horizontal spatial frequency, and the Y axis represents the vertical spatial frequency. The spatial frequency of the component close to the origin of the XY axis (the center of the spatial frequency image 500) is low, and the spatial frequency of the component away from the origin is high. The acquisition unit 302 divides the spatial frequency image 500 into a low frequency component 501, a medium frequency component 502, and a high frequency component 503 according to the distance from the origin, and sums the energy in the wedge regions 9 to 16 in the high frequency component 503. To do. The acquisition unit 302 selects image data having the maximum total energy from the image data of a plurality of frames. The spatial frequency image 500 is held in the RAM 215 even after the image data acquisition process (step S404) ends.

  The analysis unit 303 reads out the image data acquired by the acquisition unit 302 from the storage 217, and analyzes the image data. The analysis unit 303 may acquire image data to be analyzed directly from the acquisition unit 302. The analysis unit 303 performs Fourier transform on the image data, and extracts the feature amount of the image data in the frequency space (step S405). Here, the analysis unit 303 can use the spatial frequency image 500 generated by the acquisition unit 302. For example, the analysis unit 303 calculates the energy in the wedge regions 1 to 16 in the medium frequency component 502 and the high frequency component 503, respectively. As shown in FIG. 5, the wedge regions 1 to 16 include frequency components having eight different directions in the frequency space and frequency components having two different frequency bands. The analysis unit 303 uses the calculated 16 frequency energy as the feature amount of the image data. In addition, the direction and frequency band of a spatial frequency component may be selectable according to an imaging site | part, and are not limited to the example of this embodiment.

  Subsequently, the analysis unit 303 derives a cheek measurement result from the feature amount of the image data (step S405). The analysis unit 303 generates a measurement result (for example, skin moisture content, texture fineness) by applying a predetermined transformation matrix to the feature amount of the image data. The analysis unit 303 is not limited to skin, and can apply a transformation matrix adapted to the surface of the human body such as the scalp, nails, hair, and lips. These transformation matrices can be determined in advance, for example, by performing machine learning using a measurement result using a dedicated imaging device such as a microscope as teacher data. The analysis unit 303 displays the measurement result on the display 211 and ends the process of the flowchart of FIG.

  FIG. 6 is a flowchart showing details of the imaging process (step S403) according to the present embodiment. First, the control unit 301 captures one frame of image data and reads it from the frame memory 213 (step S601). The control unit 301 may trim the read image data to a predetermined size (for example, 640 × 480 pixels). Subsequently, the control unit 301 performs Fourier transform on the image data and extracts the energy of the high frequency component (step S602). The method for extracting the energy of the high frequency component is the same as the method in the image data acquisition process (step S404) described above. When the image data acquisition process (step S404) is performed at the same time, the high frequency component extraction process (step S602) is omitted.

  Next, the control unit 301 calculates the amount of change in the energy of the high frequency component between the most recently captured frames (step S603). For example, the control unit 301 calculates a difference En− (En−1) between the energy En for the latest frame n and the energy En−1 for the previous frame n−1. As the energy of the high frequency component, a moving average value of a predetermined number of frames may be used. An example of the change in energy of the high frequency component between a plurality of frames is shown in FIG.

  In FIG. 7, the vertical axis represents the energy of high frequency components, and the horizontal axis represents image data of a plurality of frames numbered in time series. That is, the frame 1 is image data of a frame first captured in imaging, and the frames 2, 3,..., Frame n are sequentially captured while the imaging continues. In the example of FIG. 7, the energy difference En− (En−1) is a positive value from frame 2 to frame 33, and is a negative value after frame 34. That is, the energy peak of the high frequency component is located in the frame 33.

  Next, the control unit 301 compares the amount of change in energy with a predetermined threshold value (step S604). For example, the control unit 301 uses 0 as a threshold value, and determines whether the energy difference En− (En−1) is 0 or more (positive or negative). When the amount of change is 0 or more (a positive value) (No in step S604), the control unit 301 continues the imaging and executes the processes in steps S601 to S604 for the next frame. When the amount of change is less than 0 (negative value) (Yes in step S604), the control unit 301 ends the imaging, and a sound effect indicating that the imaging has ended normally, “shooting is completed” Or the like is output from the speaker 208 (step S605). In the example of FIG. 7, the imaging ends after the processing of the frame 34. The control unit 301 turns off the LED 207 and returns to the processing of the flowchart of FIG.

  According to the present embodiment, by disabling the autofocus function of the imaging unit, imaging is performed for each frame with the focus distance fixed. Since the imaging is performed while moving the imaging unit so as to pass through the in-focus position, the image data of the plurality of frames obtained by imaging includes the image data focused on the imaging region as a result. . Therefore, a clear image can be easily acquired without using the autofocus function. In imaging, the amount of change in energy of the high-frequency component for each frame is monitored, and imaging is terminated when an energy peak is detected. Thereby, the user does not need to determine that the imaging unit and the imaging region are separated beyond the in-focus distance.

  The image analysis apparatus according to the present embodiment is particularly suitable for close-up imaging in which the AF function does not operate properly. By setting the in-focus distance to the shortest distance, it is possible to acquire an image in which the imaging region is clearly enlarged. Even a user who does not have a device dedicated to skin measurement can perform accurate skin measurement using a smartphone camera.

[Second Embodiment]
Next, an image analysis system according to the second embodiment of the present invention will be described. Since the image analysis system according to the present embodiment is configured in the same manner as the image analysis apparatus 100 according to the first embodiment, the description will focus on differences from the first embodiment. In the present embodiment, high frequency components of image data are extracted by Laplacian calculation.

  FIG. 8 is a diagram for explaining a high-frequency component extraction method according to the present embodiment. FIG. 8A is a conceptual diagram of processing for generating Gaussian images G0 to G2 and Laplacian images L0 to L2. In each image, circular and linear objects are shown as an example. The level 0 Gaussian image G0 is the same as the original image read from the frame memory 213.

ここで、ｘはフィルタ中心からのＸ軸（水平）方向の座標を表し、ｙはフィルタ中心からのＹ軸（垂直）方向の座標を表す。σは平滑化の度合いを決定する平滑化係数を表す。σ≒０．８におけるガウシアンフィルタの例を図８（ｂ）に示す。 First, the acquisition unit 302 performs a smoothing process on image data by applying a Gaussian filter expressed by the following equation to a level 0 Gaussian image G0.

Here, x represents the coordinate in the X-axis (horizontal) direction from the filter center, and y represents the coordinate in the Y-axis (vertical) direction from the filter center. σ represents a smoothing coefficient that determines the degree of smoothing. An example of a Gaussian filter with σ≈0.8 is shown in FIG.

  Next, the acquisition unit 302 downsamples the level 0 Gaussian image G0 to reduce the image. As a result, a level 1 Gaussian image G1 is generated. For example, the acquisition unit 302 generates a ½ size Gaussian image G1 composed of even-numbered rows and even-numbered column pixel values of the Gaussian image G0. The acquisition unit 302 further reduces the image by repeating the same processing as the level 0 Gaussian image G0 for the level 1 Gaussian image G1. As a result, a level 2 Gaussian image G2 is generated. The Gaussian images G0 to G2 are called Gaussian pyramids. The resolution of a Gaussian image decreases every time downsampling is performed. In other words, high frequency components of the image data are lost due to downsampling.

  The acquisition unit 302 upsamples the level 1 Gaussian image G1 and enlarges the image to the original size. For example, the acquisition unit 302 interpolates the odd row and odd column pixel values from the even row and even column pixel values of the Gaussian image G1, and generates an enlarged image having the same size as the level 0 Gaussian image G0. The acquisition unit 302 can acquire a level 0 Laplacian image L0 by calculating a difference between the Gaussian image G0 and the enlarged image. The level 0 Laplacian image L0 is an image of a high frequency component lost in the process of generating the Gaussian image G1. The acquisition unit 302 can acquire the level 1 Laplacian image L1 by repeating the same processing for the level 2 Gaussian image G2 and the level 1 Gaussian image G1. The level 1 Laplacian image L1 is an image of a high frequency component lost in the process of generating the Gaussian image G2. The level 1 Laplacian image L1 has a lower frequency component than the level 0 Laplacian image L0. The acquisition unit 302 may generate a level 2 Laplacian image L2 having a lower frequency component. Laplacian images L0 to L2 are called Laplacian pyramids.

  The acquisition unit 302 may generate Laplacian images L0 to L2 by applying a Laplacian filter to the above-described Gaussian images G0 to G2. A Laplacian filter is a filter that calculates a second derivative (difference) of an image. An example of a Laplacian filter in the vicinity of 8 is shown in FIG. The acquisition unit 302 multiplies neighboring pixel values centered on the target pixel by a coefficient of a Laplacian filter, and sums the multiplication results. The calculated pixel value becomes the pixel value of the Laplacian image. Note that the acquisition unit 302 is not limited to the Gaussian filter, and may use other smoothing filters such as a median filter and a moving average filter, or a LOG (Laplacian Of Gaussian) filter that combines a Gaussian filter and a Laplacian filter. May be.

  FIG. 9 is a flowchart of the image analysis method according to the present embodiment. The processing from step S901 to S903 is the same as the processing from step S401 to S403 in FIG.

  The acquisition unit 302 applies a Laplacian filter to the image data of a plurality of frames obtained in imaging, and extracts a high frequency component from the image data of each frame (step S904). For example, the acquisition unit 302 reads multiple frames of image data from the frame memory 213 and applies a Gaussian filter to the image data of each frame. As illustrated in FIG. 8, the acquisition unit 302 downsamples the image data smoothed by the Gaussian filter, and generates a level 1 Gaussian image G1. The acquisition unit 302 applies a Laplacian filter to the level 1 Gaussian image G1 to generate a level 1 Laplacian image L1.

  Next, the acquisition unit 302 acquires image data with the maximum energy of the high frequency component based on the Laplacian image L1 of each frame (step S905). Specifically, the acquisition unit 302 sums the pixel values of the Laplacian image L1 for each frame, and determines the frame that maximizes the sum of the pixel values. The acquisition unit 302 acquires the image data of the determined frame as a focused image. Note that the acquisition unit 302 may extract a high-frequency component by generating a Laplacian pyramid (Laplacian images L0, L1, and L2) illustrated in FIG. The acquisition unit 302 can determine the magnitude of the energy of the high frequency component based on any one of Laplacian images L0, L1, and L2 having different resolution levels, or a combination thereof. The subsequent processes in steps S906 and S907 are similar to those in steps S405 and S406 in FIG.

[Modified Embodiment]
The above-described embodiments are merely examples of implementation in carrying out the present invention, and the technical scope of the present invention is not construed as being limited thereto. That is, the present invention is not limited to the above-described embodiment, and can be modified without departing from the spirit of the present invention.

  For example, in the above-described embodiment, imaging is started in a state where the first distance is shorter than the second distance and the imaging unit 102 and the imaging part are closer than the in-focus distance, and the imaging unit 102 and the imaging part are A case has been described in which imaging is continued until the distance exceeds the in-focus distance. On the contrary, the imaging is started in a state where the first distance is longer than the second distance and the imaging unit 102 and the imaging part are separated from the focusing distance, and the imaging part 102 and the imaging part are aligned. Imaging may be continued until approaching within the focal distance. In other words, the image analysis device 100 is configured to be able to detect a peak of energy of a high frequency component by changing a distance between the imaging unit 102 and the imaging site for a plurality of frames of image data obtained by imaging. Good.

  Further, in the energy change amount determination process (step S604), the CPU 214 does not have to immediately finish imaging even when the energy change amount is smaller than a predetermined threshold. For example, the CPU 214 may continuously execute the processing for the next frame (steps S601 to S604), and may end imaging when the amount of change in energy is smaller than a predetermined threshold for two consecutive frames. .

  In the above-described embodiment, the case where the image data of a plurality of frames is captured by moving image capturing has been described. However, the processing of the above-described embodiment can be applied even when still images are captured continuously. Furthermore, the measurement site | part by the image analysis apparatus 100 is not restricted to a user's cheek. The image analysis apparatus 100 can be applied to measure various parts of the human body surface such as skin, scalp, nails, hair, and lips.

DESCRIPTION OF SYMBOLS 100 Image analysis apparatus 101 Touch panel 102 Imaging part 213 Frame memory 214 CPU
215 RAM
217 Storage 300 Measurement application program 301 Control unit 302 Acquisition unit 303 Analysis unit 500 Spatial frequency image 503 High frequency component

Claims (14)

An imaging unit that captures image data of a plurality of frames at a predetermined frame rate;
The focusing distance of the imaging unit is fixed by disabling the autofocus function of the imaging unit, and the imaging unit is positioned at a first distance different from the focusing distance with respect to the imaging part on the human body surface. A controller that starts the imaging in a state and continues the imaging until the imaging unit moves to a position that becomes the second distance beyond the focusing distance with respect to the imaging part;
An acquisition unit that acquires the image data having the maximum frequency component energy based on the sharpness of the imaging region from the image data of the plurality of frames;
An image analysis apparatus comprising: an analysis unit that analyzes the image data acquired by the acquisition unit.   The image analysis apparatus according to claim 1, wherein the analysis unit performs frequency conversion on the image data to calculate energy of a plurality of frequency components having different directions and frequency bands.   The image analysis apparatus according to claim 2, wherein the acquisition unit extracts the frequency component using frequency conversion, and the analysis unit analyzes the image data using a result of the frequency conversion by the acquisition unit. .   The image analysis apparatus according to claim 2 or 3, wherein the frequency transform is at least one of Fourier transform, wavelet transform, Laplacian calculation, discrete cosine transform, Hadamard transform, and KL (Karhunen-Loeve) transform. The acquisition unit calculates a change amount of the energy between image data of a predetermined number of frames captured most recently among the image data of the plurality of frames,
The image analysis apparatus according to claim 1, wherein the control unit ends the imaging when the amount of change becomes smaller than a predetermined threshold.   The image analysis apparatus according to claim 1, wherein the control unit fixes the focus distance to a shortest settable distance. It has an operation unit that accepts instructions from users,
The image analysis apparatus according to claim 1, wherein the control unit starts the imaging in response to the instruction. An illumination unit that emits light toward the imaging region;
The image analysis apparatus according to claim 1, wherein the control unit turns on the illumination unit while continuing the imaging.   The image analysis apparatus according to claim 1, further comprising a notification unit that notifies a user that the imaging has been completed.   The image analysis apparatus according to any one of claims 1 to 9, which is a smartphone, a tablet computer, a PDA (Personal Digital Assistant), a laptop computer, or a mobile phone.   The image analysis apparatus according to claim 1, wherein the first distance is shorter than the second distance.   The image analysis apparatus according to claim 1, wherein the first distance is longer than the second distance. Fixing the focusing distance of the imaging unit by disabling the autofocus function of the imaging unit;
Starting imaging a plurality of frames of image data at a predetermined frame rate in a state where the imaging unit is located at a first distance different from the in-focus distance with respect to an imaging part on the surface of a human body;
Continuing the imaging until the imaging unit moves to a position that becomes the second distance beyond the focusing distance with respect to the imaging site;
Obtaining the image data having the maximum frequency component energy based on the sharpness of the imaging region from the image data of the plurality of frames;
An image analysis method comprising: analyzing the acquired image data.   A program for causing a computer to execute each step of the image analysis method according to claim 13.

Images