Subject image extraction system, subject image extraction method, and program

Abstract

To provide a subject image extraction system, a subject image extraction method, and a program capable of capturing a subject moving at high speed and obtaining an image in which the subject is suitably captured.SOLUTION: A subject image extraction system includes: capture means that captures a series of images; moving body detection means that detects a moving body in the images by processing the respective images captured by the capture means; image storage means that, upon detecting the moving body, stores the image in which the moving body is detected; subject detection means that detects a subject from the stored images and adds an index value to each of the images based on the detected subject; and extraction image determination means that determines, based on the index value, an extracted image which is an image to be extracted based on the index value.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present disclosure relates to a subject image extraction system, a subject image extraction method, and a program. [Background technology]

[0002] Conventionally, fast-moving subjects are photographed with cameras at amusement parks, amusement parks, racing circuits, etc. Various techniques have been proposed for extracting suitable images from fast-moving subjects. Among these, a method has been disclosed in which camera footage is analyzed and detection is performed based on the movement and characteristics of people (see Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Publication No. 2019-71553 Summary of the Invention [Problem to be solved by the invention]

[0004] However, the above-mentioned conventional technology has a problem in that the equipment becomes complicated, for example, a separate sensor is required as a trigger for photographing.

[0005] Therefore, an object of the present disclosure is to provide a subject image extraction system, a subject image extraction method, and a program that are capable of photographing a subject moving at high speed and obtaining an image in which the subject is captured well. [Means for solving the problem]

[0006] In order to solve the above problems, the present disclosure provides: a photographing means for continuously photographing images; a moving object detection means for processing each image acquired by the photographing means and detecting a moving object in the image; an image storage means for storing an image of the moving object when the moving object is detected; a subject detection means for detecting a subject from the stored images and assigning an index value to each image based on the detected subject; an extraction image determination means for determining an extraction image, which is an image to be extracted, based on the index value; The present invention provides a subject image extraction system having the above structure.

[0007] In addition, in the subject image extraction system of the present disclosure, The subject detection means may process the stored image at a processing rate lower than that of the moving object detection means.

[0008] In addition, in the subject image extraction system of the present disclosure, The image storage means may successively store a number of images not exceeding a number calculated based on the subject movement distance, the subject speed, and the shooting rate.

[0009] In addition, in the subject image extraction system of the present disclosure, The moving object detection means may detect, as a moving object, a circumscribing rectangle of an area determined based on a difference between successive images.

[0010] In addition, in the subject image extraction system of the present disclosure, The moving object detection means may determine that a moving object has been detected only when the moving object detected in the set detection area is equal to or larger than a set minimum detection size.

[0011] In addition, in the subject image extraction system of the present disclosure, The detection areas may be set in a plurality of locations, and when a moving object having a minimum detection size or larger is detected in one detection area, detection processing may be performed on the other detection areas.

[0012] In addition, the subject image extraction system of the present disclosure includes: The subject detection means may detect the subject by object detection using a trained model based on machine learning, set an index value based on the state of the detected subject for each image, and determine an extracted image based on the index value.

[0013] In addition, in the subject image extraction system of the present disclosure, The index value used by the subject detection means may be a value of any one of the indexes of the position, area, and likelihood of the subject in the image.

[0014] In addition, in this disclosure, A subject image extraction method in which a computer connected to a photographing device that continuously photographs images extracts images in which a subject appears as extracted images, comprising: The computer processing each image acquired by the image capture device to detect moving objects in the image; When the moving object is detected, an image of the moving object is stored; Detecting a subject from the stored images, and assigning an index value to each image based on the detected subject; determining an extraction image, which is an image to be extracted, based on the index value; The present invention provides a method for extracting an object image.

[0015] In addition, in this disclosure, A computer connected to an imaging device that captures an image of a subject, a moving object detection means for processing each image acquired by the photographing device and detecting a moving object in the image; an image storage means for storing an image of the moving object when the moving object is detected; a subject detection means for detecting a subject from the stored images and assigning an index value to each image based on the detected subject; an extraction image determination means for determining an extraction image, which is an image to be extracted, based on the index value; We provide a program to function as a [Effects of the Invention]

[0016] According to the present disclosure, it is possible to provide a subject image extraction system, a subject image extraction method, and a program that are capable of capturing an image of a subject moving at high speed and obtaining a suitable still image. [Brief explanation of the drawings]

[0017] [Figure 1] 1 is a diagram illustrating a configuration of a subject image extraction system according to an embodiment of the present disclosure. [Figure 2] FIG. 2 is a functional block diagram showing details of the processing device 11. [Figure 3] FIG. 1 is a state transition diagram illustrating an overview of processing of a subject image extraction system according to an embodiment of the present disclosure. [Figure 4] FIG. 2 is a diagram showing the processing rate in each means (each state) in this embodiment. [Figure 5] 10A and 10B are diagrams for explaining the moving object detection process performed by the moving object detection means 11a. [Figure 6] 10A and 10B are diagrams for explaining another method of moving object detection processing by the moving object detection means 11a. [Figure 7] 11A and 11B are diagrams for explaining the subject detection process performed by the subject detection means 11d. [Figure 8] 11A and 11B are diagrams for explaining the extraction image determination process performed by the extraction image determination means 11e. DETAILED DESCRIPTION OF THE INVENTION

[0018] Preferred embodiments of the present disclosure will be described in detail below with reference to the drawings. <System configuration> Fig. 1 is a diagram showing the configuration of a subject image extraction system according to an embodiment of the present disclosure. In Fig. 1, 10 denotes a subject image extraction system, 11 denotes a processing device, 12 denotes a photographing device, 13 denotes a display device, 14 denotes a printer, and 15 denotes a video data output unit.

[0019] The processing device 11 performs various arithmetic operations in the subject image extraction system 10 and is realized by incorporating a program for implementing the functions of the subject image extraction system 10 into a computer having a CPU, main memory, non-volatile storage device, etc. The processing device 11 is connected to the photographing device 12. The photographing device 12 is a photographing means that captures photographed images as image data and is realized by a video camera equipped with an image sensor such as a CCD. A video is obtained by continuously capturing still images. In this specification, the term "image" simply refers to a still image. The display device 13 is a display means that displays the results processed by the processing device 11 and is realized by a general-purpose display device such as an LCD display or an organic EL display. The printer 14 and the video data output unit 15 are image output means that output images. The printer 14 prints images on paper or other media. The video data output unit 15 is realized by a data output device that outputs a series of images as a video.

[0020] 2 is a functional block diagram showing details of the processing device 11. In this embodiment, the processing device 11 is realized by installing a dedicated program in a general-purpose computer. As shown in FIG. 2, the processing device 11 includes a moving object detection unit 11a, an image saving unit 11b, an image storage unit 11c, a subject detection unit 11d, and an extracted image determination unit 11e.

[0021] The moving object detection means 11a is a means for performing image analysis on the photographed image acquired by the photographing device 12 and detecting a moving object in the photographed image, and is realized by the CPU reading and executing a program stored in a nonvolatile storage device. A moving object means an object that moves. The image saving means 11b is a means for saving an image in which a moving object is detected in the image storage means 11c, when a moving object is detected, and is realized by the CPU reading and executing a program stored in a nonvolatile storage device.

[0022] The image storage means 11c is a storage means for storing images stored by the image storage means 11b. The image storage means 11c is realized by a non-volatile storage device. The subject detection means 11d is a means for detecting subjects from the stored images and assigning an index value to each image based on the detected subjects, and is realized by the CPU reading and executing a program stored in the non-volatile storage device. The extracted image determination means 11e is a means for determining an extracted image, which is an image to be extracted, based on the index value, and is realized by the CPU reading and executing a program stored in the non-volatile storage device.

[0023] The processing device 11 is provided with an instruction input unit such as a keyboard and a mouse, and receives instructions from a user. Various setting values ​​can be set based on the instructions from the user.

[0024] <Processing operation> Next, the processing operation of the subject image extraction system according to this embodiment will be described together with the subject image extraction method according to this embodiment. Figure 3 is a state transition diagram showing an outline of the processing of the subject image extraction system according to this embodiment. After being started, the subject image extraction system according to this embodiment performs processing while transitioning through four states S1 to S4.

[0025] First, when the subject image extraction system is started, processing begins. Then, the photographing device 12 constantly takes photographs and captures the image data obtained by photographing as photographed images. The captured photographed images are passed to the processing device 11 for predetermined processing. The photographing device 12 takes photographs at a predetermined rate (fps: frames (images) per second), captures the captured images, and sends them to the processing device 11. The predetermined rate can be set arbitrarily. In this embodiment, it is set to 60 fps. The photographing device 12 always acquires photographed images while the subject image extraction system is operating, regardless of the states S1 to S4 shown in FIG. 3.

[0026] When processing begins in the subject image extraction system, the system first transitions to a moving object detection state S1. In the moving object detection state S1, the image capture device 12 continuously captures images and captures the image data obtained by the capture as a captured image. The processing device 11 then performs processing such as analysis on the captured image. The analysis rate of the captured image in the processing device 11 may be the same as the capture rate of the image capture device 12, or it may be different. For example, if the capture rate is 60 fps as described above, the analysis rate in the processing device 11 may be the same as the capture rate, 60 fps, or may be a slower rate, for example, 40 fps.

[0027] In the moving object detection state S1, the moving object detection means 11a in the processing device 11 performs moving object detection processing. Then, when the moving object detection means 11a detects moving objects of a predetermined size or more consecutively in one to several areas, the state transitions to the image storage state S2. If no person is detected in the captured image, the moving object detection means 11a repeatedly performs moving object detection processing on newly captured images captured by the image capture device 12. Details of the moving object detection processing will be described later.

[0028] In the image saving state S2, the image saving means 11b continues to save the captured images in the image storage means 11c after the moving object detection means 11a detects a moving object. Then, when the number of captured images saved in the image storage means 11c reaches a predetermined number, the state transitions to the subject detection state S3.

[0029] In the subject detection state S3, the subject detection means 11d detects subjects and assigns index values ​​to the images stored in the image storage means 11c. Details of subject detection and assignment of index values ​​by the subject detection means 11d will be described later. If one or more subjects are detected, the state transitions to the extracted image determination state S4. If no subjects are detected, the state transitions to the moving object detection state S1.

[0030] In the extracted image determination state S4, the extracted image determination means 11e determines extracted images that should be extracted. Images that should be extracted are images in which the subject is clearly visible. The extracted image determination means 11e detects the subject from the image, assigns an index value to the image, and determines a specified number of extracted images based on the index value. Once the extracted image determination means 11e has determined the extracted images, the processing device 11 displays the extracted images on the display device 13. The user can check the extracted images displayed on the display device 13. When an output instruction is given to the processing device 11 by a user's operation, the processing device 11 outputs the extracted images to the image output means. If the image output means is the printer 14, each image is printed out. If the image output means is the video data output unit 15, multiple images are output as video data that can be displayed continuously. After completing the processing in the extracted image determination state S4, the processing device 11 transitions to the moving object detection state S1.

[0031] <Processing rate for each state> In this embodiment, the processing rate differs among the states S1 to S4. In the moving object detection state S1 and the image storage state S2, all captured images must be processed in real time at the same rate as the capture rate. That is, the moving object detection means 11a and the image storage means 11b perform processing at the same processing rate as the capture rate of the image capture device 12. On the other hand, the subject detection state S3 processes a certain number of stored images, and does not need to process at the same rate as the capture rate. Furthermore, the extracted image determination state S4 processes index values ​​assigned to each image, and does not need to process at the same rate as the capture rate. That is, the subject detection means 11d and the extracted image determination means 11e may perform processing at a processing rate lower than the capture rate of the image capture device 12.

[0032] FIG. 4 is a diagram showing the processing rates of each means (each state) in this embodiment. FIG. 4 shows a case where the shooting rate of the shooting device 12 is 60 fps. As shown in FIG. 4, in this embodiment, the processing rates of the moving object detection means 11a and the image storage means 11b are both 60 fps. Furthermore, the processing rates of the subject detection means 11d and the extracted image determination means 11e are both 10 fps. In this way, by keeping the processing rates of the subject detection means 11d and the extracted image determination means 11e lower than the shooting rate, the processing load on the processing device 11 can be reduced.

[0033] <Setting the number of images to save> The number of images to be saved by the image saving means 11b when the moving object detecting means 11a detects a moving object is calculated by the following (Equation 1) and (Equation 2).

[0034] Number of images saved = Subject movement time × Shooting rate (Formula 1) Photography subject movement time = Photography subject movement distance ÷ Photography subject speed (Formula 2)

[0035] The shooting rate in (Equation 1) is determined by the specifications and settings of the shooting device 12, such as a video camera. The subject movement time in (Equation 1) is calculated using (Equation 2). The subject speed in (Equation 2) can be set in advance. For example, the speed of a roller coaster at an amusement park or a race car at a racing track at a predetermined position visible from the shooting point is set within an approximate range. Therefore, the expected subject speed can be set in advance.

[0036] The distance traveled by the subject in equation (2) is determined by the preset installation location of the image capture device 12 and the expected range of movement of the moving object. For example, assume that the angle of view of the image capture device 12 installed to capture a moving object is 10 m wide x 6 m high, calculated from the distance to the moving object. In this case, the maximum diagonal movement distance is 11.7 m. In this case, it can be rounded up to 12 m.

[0037] For example, the set number of sheets to be saved is calculated under the following conditions. Subject movement distance = 12m Subject speed = 60km / h Shooting rate = 60fps

[0038] In the above conditions, Subject movement time = 0.72 seconds (Equation 2) Number of saved settings = 43 (from formula 1)

[0039] The set number of images to be saved is calculated by rounding down to the nearest whole number. In this way, the set number of images to be saved can be determined based on the shooting rate, the subject speed (speed of the moving object), the assumed position in the angle of view of the image capturing device 12, etc. The determined shooting rate, subject speed, etc. are input by the instruction input unit and set by the processing device 11. The image saving means 11b will continuously save a number of images that does not exceed the number of images to be saved calculated based on the subject movement distance, subject speed, and shooting rate.

[0040] <Motion detection> The processing of the moving object detection means 11a in the moving object detection state S1 will be described. FIG. 5 is a diagram for explaining the moving object detection processing by the moving object detection means 11a. As described above, the moving object detection means 11a processes captured images at the same processing rate as the shooting rate. The moving object detection means 11a extracts a moving object region using two consecutive captured images. FIG. 5(a) is an image captured at time t, and FIG. 5(b) is an image captured at time t+1. The image captured at time t and the image captured at time t+1 are two consecutive captured images.

[0041] The moving object detection means 11a extracts a moving object region by comparing and analyzing two or more captured images. Specifically, the moving object region is extracted using the difference between consecutive images. In the example of FIG. 5, two images are used. In FIGS. 5(a) and 5(b), only the car is moving. Therefore, as shown in FIG. 5(c), the area surrounded by the outline of the car after it has moved is extracted as the moving object region. Furthermore, the moving object detection means 11a obtains a circumscribing rectangle for the extracted moving object region and detects this circumscribing rectangle as the moving object. As shown in FIG. 5(d), by identifying the circumscribing rectangle as the moving object, the computational load in subsequent processing can be reduced.

[0042] The moving object detection means 11a performs high-speed processing at the same processing rate as the shooting rate. Therefore, detection accuracy is relatively low, and objects other than the intended object may be detected. For example, shadows, sudden weather changes, flying objects, birds, etc. may be erroneously detected. To correct such erroneous detection, in this embodiment, the object detection means 11d performs detailed detection in the subsequent object detection state S3. Conversely, because the object detection means 11d performs detailed detection, the moving object detection means 11a can perform high-speed processing in accordance with the shooting rate.

[0043] The moving object detection means 11a can perform moving object detection using various methods other than the example shown in FIG. 5. FIG. 6 is a diagram for explaining another method of moving object detection processing by the moving object detection means 11a. FIG. 6(a) shows one captured image, and FIG. 6(b) shows a state in which a detection area has been set in the captured image shown in FIG. 6(a). The example in FIG. 6 is a method suitable for cases in which the route of a moving object is specified and the moving speed of the moving object is fast. FIG. 6 shows a roller coaster in which the moving route is specified and the moving speed of the moving object is fast. In such a case, since the moving route is determined on a rail, a detection area is set on the rail in advance. As long as there is more than one detection area, any number of locations can be set.

[0044] In the example of FIG. 6(b), two detection areas are set. When multiple detection areas are set like this, the moving object detection means 11a first performs detection processing on the first detection area. Then, as shown in FIG. 5, comparison is performed only on the detection areas in two consecutive captured images. To improve detection accuracy, three or more images may be used. FIG. 6(c) is a diagram showing an expanded detection area. As shown in FIG. 6(c), a minimum detection size is set for each detection area. The moving object detection means 11a determines that a moving object has been detected only if the moving object detected in the set detection area is equal to or larger than the set minimum detection size. If a moving object equal to or larger than the minimum detection size is detected in the first detection area, the moving object detection means 11a performs detection processing on the second detection area. If a moving object equal to or larger than the minimum detection size is also detected in the second detection area, the moving object detection means 11a determines that a moving object has been detected. By comparing only the pre-set detection areas as shown in FIG. 6, processing load can be reduced and processing can be performed faster than when the entire captured image is compared as shown in FIG. 5.

[0045] <Subject detection> The processing of the subject detection means 11d in the subject detection state S3 will be described. FIG. 7 is a diagram for explaining the subject detection processing by the subject detection means 11d. FIG. 7 shows a sequence of images from top to bottom. Of these, three images G1, G2, and G3 are shown in FIG. 7. As described above, the subject detection means 11d processes the saved images at a processing rate lower than the shooting rate. The subject detection means 11d analyzes the saved images to detect subjects. The analysis of the captured images may use a trained model for object detection using machine learning, or may use pattern matching using a subject pattern prepared in advance.

[0046] Once a subject is detected, the subject detection means 11d identifies the position of the detected subject on the captured image. Specifically, the subject position and subject area within the captured image are set. The subject position is expressed as the distance from the center pixel of the captured image to the center pixel of the subject. Therefore, if the subject position is 0, it means that the subject is at the center of the image. The closer the subject position is to 0, the closer the subject is to the center of the image, and the larger the value of the subject position, the farther the subject is from the center of the image.

[0047] The subject area indicates the area of ​​the subject in the photographed image. The subject area is set in proportion to the number of pixels in the photographed image. In this way, the subject detection means 11d assigns an index value to each photographed image stored in the image storage means 11c.

[0048] In image G1, the subject is relatively small, so the subject area is 2, in image G2, the person is medium-sized, so the subject area is 6, and in image G3, the subject is relatively large, so the subject area is 8. In addition, in image G1, the subject is relatively far from the center, so the subject position is 8, in image G2, the subject is located at the center, so the subject position is 0, and in image G3, the subject is relatively close to the center, so the subject position is 6.

[0049] The subject area may be any index that can express the relative size of the subject, and is not limited to the example in Fig. 7. Furthermore, the subject position does not necessarily have to be the distance from the center, and may be any index that can relatively express the position in the captured image.

[0050] <Extracted image selection> The processing of the extraction image determination means 11e in the extraction image determination state S4 will now be described. FIG. 8 is a diagram for explaining the extraction image determination processing by the extraction image determination means 11e. As described above, the extraction image determination means 11e processes the captured image at a processing rate lower than the shooting rate. In practice, it is preferable to process at the same processing rate as the subject detection means 11d. In the extraction image determination state S4, an extraction image is determined based on the index value of the subject detected by the subject detection means 11d.

[0051] The index value of the subject to be used to determine the extracted images can be set in advance. For example, if priority is given to the index of the subject position (position priority), a predetermined number of images are extracted starting from the image whose subject position is closest to the center. For example, if priority is given to the index of the subject area (area priority), a predetermined number of images are extracted starting from the image whose subject area is largest. The number of images to be extracted may be set in advance, or only the number whose index value satisfies a predetermined condition may be extracted.

[0052] For example, when capturing images at high speeds such as 60 fps, the position and size (area) of the subject in the captured image gradually change. Therefore, whether position priority is selected or area priority is selected, consecutive images are often extracted. If the setting is to extract 10 images with position priority, 10 consecutive images including image G2 are extracted as shown in Figure 8. If the setting is to extract 10 images with area priority, 10 consecutive images including image G3 are extracted as shown in Figure 8. In addition to the position and area of ​​the subject, likelihood (probability) may also be used as an index to determine the extracted images. Likelihood indicates the degree of possibility that the subject is the subject, and can be a score obtained during object detection using a trained model.

[0053] In this manner, the extracted photographed image is output to an image output unit such as printer 14 or video data output unit 15. The image output to the image output unit may be the extracted image determined by extracted image determination means 11e directly. Alternatively, the extracted image determined by extracted image determination means 11e may be displayed on display device 13, and after the content is confirmed, the image may be output to the image output unit based on an instruction from the instruction input unit by the user.

[0054] While the preferred embodiments of the present disclosure have been described above, the present disclosure is not limited to the above embodiments and various modifications are possible. For example, in the above embodiments, a computer is used as the processing device 11, and a CPU executes a dedicated program stored in a storage device to realize a moving object detection means, an image storage means, a subject detection means, and an extracted image determination means. However, each of these means may be incorporated into hardware as an arithmetic circuit. [Explanation of symbols]

[0055] 11 Processing equipment 11a...Moving object detection means 11b...Image storage means 11c...Image storage means 11d. Subject detection means 11e Extracted image determination means 12. Imaging device 13...Display device 14. Printer 15. Video data output section

Claims

1. a photographing means for continuously photographing images; a moving object detection means for processing each image acquired by the photographing means and detecting a moving object in the image; an image storage means for storing an image of the moving object when the moving object is detected; a subject detection means for detecting a subject from the stored images and assigning an index value to each image based on the detected subject; an extraction image determination means for determining an extraction image, which is an image to be extracted, based on the index value; and The subject detection means processes the stored image at a processing rate lower than that of the moving object detection means.

2. a photographing means for continuously photographing images; a moving object detection means for processing each image acquired by the photographing means and detecting a moving object in the image; an image storage means for storing an image of the moving object when the moving object is detected; a subject detection means for detecting a subject from the stored images and assigning an index value to each image based on the detected subject; an extraction image determination means for determining an extraction image, which is an image to be extracted, based on the index value; and The image storage means continuously stores a number of images not exceeding a number calculated based on the subject movement distance, the subject speed, and the shooting rate.

3. 3. The subject image extraction system according to claim 1, wherein said moving object detection means detects, as a moving object, a circumscribing rectangle of an area determined based on a difference between successive images.

4. 4. The subject image extraction system according to claim 1, wherein the moving object detection means determines that a moving object has been detected only if the moving object detected in the set detection area is equal to or larger than a set minimum detection size.

5. 5. The subject image extraction system according to claim 4, wherein the detection areas are set in a plurality of locations, and when a moving object having a minimum detection size or larger is detected in one detection area, detection processing is performed on the other detection areas.

6. 6. The subject image extraction system according to claim 1, wherein the subject detection means detects the subject by object detection using a trained model based on machine learning, sets an index value based on the state of the detected subject for each image, and determines an extracted image based on the index value.

7. 7. The object image extraction system according to claim 1, wherein the index value used by the object detection means is a value of an index of one of a position, an area, and a likelihood of the object in the image.

8. A subject image extraction method in which a computer connected to a photographing device that continuously photographs images extracts images in which a subject appears as extracted images, comprising: The computer processing each image acquired by the image capture device to detect moving objects in the image; When the moving object is detected, an image of the moving object is stored; Detecting a subject from the stored images, and assigning an index value to each image based on the detected subject; determining an extraction image, which is an image to be extracted, based on the index value; The subject image extraction method comprises processing the stored image at a processing rate lower than that of the moving object detection.

9. A subject image extraction method in which a computer connected to a photographing device that continuously photographs images extracts images in which a subject appears as extracted images, comprising: The computer processing each image acquired by the image capture device to detect moving objects in the image; When the moving object is detected, an image of the moving object is stored; Detecting a subject from the stored images, and assigning an index value to each image based on the detected subject; determining an extraction image, which is an image to be extracted, based on the index value; The subject image extraction method includes continuously saving the images in a number not exceeding the number of images to be saved calculated based on the subject movement distance, the subject speed, and the shooting rate.

10. A computer connected to an imaging device that captures an image of a subject, a moving object detection means for processing each image acquired by the photographing device and detecting a moving object in the image; an image storage means for storing an image of the moving object when the moving object is detected; a subject detection means for detecting a subject from the stored images and assigning an index value to each image based on the detected subject; an extraction image determination means for determining an extraction image, which is an image to be extracted, based on the index value; It functions as The subject detection means processes the stored image at a processing rate lower than that of the moving object detection means.

11. A computer connected to an imaging device that captures an image of a subject, a moving object detection means for processing each image acquired by the photographing device and detecting a moving object in the image; an image storage means for storing an image of the moving object when the moving object is detected; a subject detection means for detecting a subject from the stored images and assigning an index value to each image based on the detected subject; an extraction image determination means for determining an extraction image, which is an image to be extracted, based on the index value; It functions as The image saving means is a program for continuously saving a number of images not exceeding a number calculated based on the subject movement distance, the subject speed, and the shooting rate.

Images