Information processing apparatus, monitoring apparatus, and information processing method

Abstract

To divide an image to be input to an object detection model so as to appropriately detect an object.SOLUTION: An information processing device 1 includes an acquisition unit 131 that acquires moving image data captured by an imaging device, a detection unit 132 that detects a predetermined subject in a frame image included in the moving image data, and a determination unit 133 that determines a plurality of divided regions generated by dividing an image region of the frame image and an overlap region in which each of the plurality of divided regions overlaps another adjacent divided region on the basis of a size of a region in which the predetermined subject is detected.SELECTED DRAWING: Figure 3

Description

[Technical Field]

[0001] The present invention relates to an information processing device, a monitoring device, and an information processing method. [Background technology]

[0002] A technology for performing object detection using deep learning is known (see, for example, Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2023-161956 Summary of the Invention [Problem to be solved by the invention]

[0004] However, when there is a limit to the amount of information that can be input into a trained model that performs image recognition, the input image must be divided, which creates the problem of reduced object detection accuracy near the boundaries of the divided images.

[0005] The present invention has been made in consideration of these points, and aims to make it possible to segment an image to be input to an object detection model so that an object can be detected appropriately. [Means for solving the problem]

[0006] An information processing device of the first aspect of the present invention includes an acquisition unit that acquires moving image data captured by an imaging device, a detection unit that detects a predetermined subject in a frame image included in the moving image data, and a determination unit that determines a plurality of divided areas generated by dividing the image area of ​​the frame image based on the size of the area in which the predetermined subject is detected, and an overlap area in which each of the plurality of divided areas overlaps with other adjacent divided areas.

[0007] The detection unit may detect the specified subject contained in each different position in the frame image, and the determination unit may determine the multiple divided areas corresponding to the position and the overlap area based on the position at which the specified subject is detected and the size of the area at that position at which the specified subject is detected.

[0008] The image processing device may further include a memory unit that stores size information indicating the size of the specified subject, and the determination unit may correct at least one of the divided area and the overlap area based on the size of the specified subject indicated by the size information.

[0009] The detection unit may further detect the position of an obstacle that is a subject included in a frame that constitutes the video image data so as to obscure at least a portion of the specified subject, and the determination unit may determine the divided area and the overlap area based on the position of the obstacle detected by the detection unit.

[0010] The apparatus may further include a calculation unit that calculates the movement speed of the specified subject detected by the detection unit, and the determination unit may determine the divided area and the overlapping area so that the larger the movement speed of the specified subject, the larger at least one of the divided area and the overlapping area becomes.

[0011] The device may further include a generation unit that generates a plurality of split frame images, which are image data obtained by dividing frames included in the target video data, based on the split area and the overlap area determined by the determination unit and the target video data, which is a video image of the monitored subject captured by the imaging device, and the detection unit may detect the target subject included in the plurality of split frames by inputting image data having a lower resolution than the frames included in the target video data into a trained model that is trained to detect target subjects, which are subjects of the monitored subject included in the image data.

[0012] A monitoring device according to a second aspect of the present invention includes an imaging unit, an acquisition unit that acquires moving image data captured by the imaging unit, a detection unit that detects a predetermined subject in a frame image included in the moving image data, and a determination unit that determines a plurality of divided areas generated by dividing the image area of ​​the frame image based on the size of the area in which the predetermined subject is detected, and an overlap area in which each of the plurality of divided areas overlaps with other adjacent divided areas.

[0013] An information processing method of the third aspect of the present invention includes the steps of: acquiring moving image data captured by an imaging device; detecting a predetermined subject in a frame image included in the moving image data; and determining, based on the size of the area in which the predetermined subject is detected, a plurality of divided areas generated by dividing the image area of ​​the frame image, and overlapping areas in which each of the plurality of divided areas overlaps with other adjacent divided areas. [Effects of the Invention]

[0014] According to the present invention, it is possible to achieve the effect of dividing an image to be input to an object detection model so that an object can be appropriately detected. [Brief explanation of the drawings]

[0015] [Figure 1] FIG. 1 is a diagram illustrating an overview of an information processing system S according to a first embodiment. [Figure 2] FIG. 1 is a diagram illustrating an overview of an information processing system S according to a first embodiment. [Figure 3] 1 is a block diagram showing a configuration of an information processing device 1. FIG. [Figure 4] FIG. 2 is a diagram for explaining division processing of the information processing device 1. [Figure 5] FIG. 2 is a diagram for explaining division processing of the information processing device 1. [Figure 6] 10 is a diagram showing an example of the data structure of division parameters generated by a determination unit 133. FIG. [Figure 7]3 is a flowchart showing the flow of processing in the information processing device 1. [Figure 8] FIG. 2 is a block diagram showing the configuration of a monitoring device 3. DETAILED DESCRIPTION OF THE INVENTION

[0016] First Embodiment [Outline of Information Processing System S] 1 and 2 are diagrams illustrating an overview of an information processing system S according to a first embodiment. The information processing system S is a system for monitoring moving image data captured by a surveillance camera or the like. The information processing system S includes an information processing device 1 and an imaging device 2. The information processing device 1 and the imaging device 2 are connected to each other so as to be able to communicate with each other via a network.

[0017] The information processing device 1 is a device for analyzing moving images captured by the imaging device 2. For example, the information processing device 1 detects objects included in the moving image data captured by the imaging device 2, and performs processing such as issuing an alarm when a specific object is detected.

[0018] The imaging device 2 is, for example, a surveillance camera. The imaging device 2 transmits captured video data to the information processing device 1.

[0019] When the resolution of moving image data captured by the imaging device 2 exceeds the resolution that can be processed by the information processing device 1, the information processing device 1 analyzes the moving image data by dividing the frame images that make up the moving image data captured by the imaging device 2. Here, the information processing device 1 needs to divide the images so that objects near the boundaries of the divided frame images can be properly detected.

[0020] FIG. 2 is a diagram for explaining the division process of frame images in the information processing system S. The information processing device 1 acquires moving image data captured by the imaging device 2. The moving image data is composed of a plurality of frame images. The information processing device 1 detects a predetermined subject included in the frame images that make up the acquired moving image data. One example of the predetermined subject is a person, but is not limited to this. The predetermined subject may be an animal such as a dog or cat, or may be a car, building, structure, etc. The method by which the information processing device 1 detects the predetermined subject will be described later. The information processing device 1 sets a rectangular area (hereinafter referred to as a "rectangular area") B in the area in the frame image P where the predetermined subject is detected.

[0021] Based on the size of area B, the information processing device 1 determines divided areas D (D1, D2, D3, and D4) into which the frame image P is divided, and an overlap area R. The divided areas D are areas into which the frame image is divided. The overlap area R is an area that is set to prevent erroneous recognition of a predetermined subject located near the boundary of the divided areas D when the image is divided. The overlap area R is set so that adjacent divided areas D include the overlap area R.

[0022] It is preferable that the width and height of the overlap region R are set to be the same as the width and height of the rectangular region B. That is, the width W of the overlap region R is R and height H R are the width W of the rectangular region B, respectively. B and height H RBy setting it in this way, even if there is a rectangular area B in the middle position of multiple divided areas D (i.e., the center position of the overlap area R), the entire rectangular area B will be included in the overlap area R, and therefore the entire rectangular area B will be included in at least one of the divided areas D. As a result, it is possible to prevent subjects near the boundaries of multiple divided areas D from going undetected. However, the above is merely a suitable example, and the width and height of the overlap area R may be increased or decreased within a range that does not affect the recognition accuracy of objects near the boundaries of the divided areas D.

[0023] The size of the divided region D is set to be larger than the size of region B and to allow the frame image to be divided into an integer number of regions. As an example, the width and height of the divided region D are an integer multiple (e.g., three times) of the width and height of the rectangular region B, but are not limited to this.

[0024] In the above, an example has been described in which all divided regions D are the same size, but this is not limiting. As will be described later, the size of divided region D may change depending on its vertical position in the frame image, its horizontal position, or both.

[0025] Although details will be described later, the information processing device 1 stores division parameters, which are information for specifying the determined divided regions D.

[0026] The information processing device 1 determines a plurality of divided areas D based on the size of a predetermined subject detected during an initial setup period after the installation of the imaging device 2, for example. Based on the determined divided areas D, the information processing device 1 acquires video data captured during monitoring, divides frame images constituting the acquired video data, and detects a predetermined subject included in the frame image to be determined.

[0027] By configuring the information processing device 1 in this manner, a subject to be monitored that is located near the boundary of the divided frame can be detected in one of the divided areas, and the image can be divided so that an object contained in the video image data can be properly detected.

[0028] [Configuration of information processing device 1] 3 is a block diagram showing the configuration of the information processing device 1. The information processing device 1 has a communication unit 11, a storage unit 12, and a control unit 13. The control unit 13 has an acquisition unit 131, a detection unit 132, a determination unit 133, a calculation unit 134, and a generation unit 135.

[0029] The communication unit 11 is a communication interface for transmitting and receiving data to and from other devices via a network. The storage unit 12 is a storage medium including a ROM (Read Only Memory), a RAM (Random Access Memory), an SSD (Solid State Drive), a hard disk drive, etc. The storage unit 12 stores in advance a program to be executed by the control unit 13.

[0030] The control unit 13 is a processor such as a CPU (Central Processing Unit), etc. The control unit 13 executes a program stored in the storage unit 12, thereby functioning as an acquisition unit 131, a detection unit 132, a determination unit 133, a calculation unit 134, and a generation unit 135.

[0031] The acquisition unit 131 acquires video data captured by the imaging device 2. The detection unit 132 detects a predetermined subject in a frame image included in the video data. The detection unit 132 inputs the frame images constituting the video data into a detection model, which is a trained model trained to detect a predetermined subject included in the input image, and detects the predetermined subject. As an example, the detection model receives image data as input and outputs information indicating an area in the input image data where the predetermined subject is included. Note that if the number of pixels in the frame image or divided region D does not match the number of dimensions of the input layer of the trained model, the detection unit 132 may input an image obtained by resizing the frame image or divided region D to the detection model.

[0032] Based on the size of area B where a predetermined subject is detected, determination unit 133 determines a plurality of divided areas D generated by dividing the image area of ​​the frame image, and overlap areas R where each of the plurality of divided areas D overlaps with adjacent divided areas D. The process of determining divided areas D and overlap areas R by determination unit 133 is as described above.

[0033] Surveillance cameras are generally installed so that they capture the area being monitored from a high position looking down. In the video captured by a surveillance camera installed in this manner, the higher an object appears in the image, the smaller it appears than its actual size. Therefore, the information processing device 1 may be configured so that the sizes of the divided area D and the overlap area R differ depending on the position where a specific object is detected in the image.

[0034] Therefore, the detection unit 132 detects a predetermined subject that is included in each of the frame images at different positions. As an example, the detection unit 132 may detect a predetermined subject from each of multiple frame images in the same video data, or may detect multiple predetermined subjects from a single frame image.

[0035] Fig. 4 is a diagram showing an example of processing when detection unit 132 detects a predetermined subject from each of a plurality of frame images. In Fig. 4, the predetermined subject is a person, who moves while changing position within the area that imaging device 2 is to capture. In Fig. 4, detection unit 132 detects the predetermined subject in frame image P1 and frame image P2, and sets rectangular area B1 and rectangular area B2 for each frame image. Frame image P2 is a frame image taken at a later time than frame image P1.

[0036] The determination unit 133 determines a plurality of divided areas D corresponding to a position where a predetermined subject has been detected and an overlapping area R based on the position and the size of the area where the predetermined subject has been detected at that position. First, the determination unit 133 determines the divided area D and overlapping area R at that position based on the position and size of a rectangular area B1 in the frame image P1. As an example, the determination unit 133 determines the width and height of the overlapping area R in a divided area that includes the horizontal position of rectangular area B1 based on the width and height of rectangular area B. Then, the determination unit 133 determines divided areas (D11, D12, and D13) that include the vertical position of rectangular area B1 so that adjacent divided areas include the same overlapping area.

[0037] Furthermore, the determination unit 133 identifies an undivided region. An undivided region is a region of the image region included in the frame image that does not belong to any of the divided regions D. If there is no undivided region, the determination unit 133 ends the division process. If there is an undivided region, the determination unit 133 performs further division processing on the identified undivided region and an overlap region R that borders the undivided region.

[0038] Similarly, based on the position and size of rectangular area B2 for frame image P2, determination unit 133 determines divided areas (D21, D22, D23, and D24) that include the vertical position of rectangular area B2 and overlap area R. Determination unit 133 repeats the above process until no undivided areas remain.

[0039] In general, there is little change in the position of a predetermined subject between temporally adjacent frames. Therefore, determining the divided area D and overlap area R for all frame images places a heavy load on the information processing device 1. Therefore, the determination unit 133 may extract frame images that satisfy predetermined conditions, and, based on a rectangular area B in the frame image, determine the divided area D and overlap area R that correspond to the position of the rectangular area B.

[0040] One example of the predetermined condition is that rectangular area B is included in the undivided area. Another predetermined condition is that rectangular area B output by detection unit 132 is enlarged or reduced by a predetermined threshold or more from the size of rectangular area B in the frame image already used to determine divided area D. Another predetermined condition may be that rectangular area B output by detection unit 132 is farther away from the position of rectangular area B in the frame image already used to determine divided area D than a predetermined threshold or more.

[0041] In the above description, divided areas are set based on predetermined subjects that appear in multiple frame images, but the information processing device 1 may be configured to set divided areas of different sizes based on multiple predetermined subjects that appear in a single frame image. When multiple predetermined subjects are detected from a single frame image, the frame image contains multiple subjects of the same type as the predetermined subjects.

[0042] 5 is a diagram for explaining the processing of the information processing device 1 in this case. The detection unit 132 detects a predetermined number of subjects included in the frame image P and sets rectangular area B11 and rectangular area B12. The determination unit 133 identifies rectangular area B11 as the largest rectangular area B. As already explained, the determination unit 133 determines divided areas (D11, D12, and D13) and overlap area R at the position of rectangular area B11 based on the position and size of rectangular area B11.

[0043] Next, the determination unit 133 identifies the largest rectangular area B among the rectangular areas B included in the undivided area as the rectangular area B to be processed. In Fig. 5, the determination unit 133 identifies rectangular area B12 as the rectangular area B to be processed. Based on the position and size of rectangular area B12, the determination unit 133 determines divided areas (D21, D22, D23, and D24) and overlap area R at the position of rectangular area B12.

[0044] If an image area that does not belong to any of the divided areas D occurs after all of the rectangular areas B have been processed, the determination unit 133 may set a divided area D of the same size as the divided area D that is closest to that area. Furthermore, the determination unit 133 may determine the size of the divided area D to be set for the area that does not belong to any of the divided areas D, based on the average size of the divided areas D that sandwich the area that does not belong to any of the divided areas D.

[0045] In the above example, the divided areas B and the overlap areas R are determined in descending order of the size of the rectangular areas B, but this is not limiting. For example, the processing may be performed in ascending order of the vertical positions of the rectangular areas B.

[0046] The determination unit 133 generates division parameters based on the determined divided region D and overlap region R. FIG. 6 is a diagram showing an example of the data structure of the division parameters generated by the determination unit 133. The division parameters include an "area ID," "reference point coordinates," and "size." The "area ID" is an ID for identifying the divided region D. The "reference point coordinates" indicate the coordinates of a position that serves as a reference for the divided region D. The "reference point coordinates" may indicate the top left position of the divided region D or may indicate the center of the divided region D. The "size" is information that indicates the size of the divided region D. As an example, the "size" indicates the number of pixels included in the divided region D. The "size" includes information that indicates the width and height of the divided region D. The determination unit 133 stores the generated division parameters in the storage unit 12.

[0047] By having the information processing device 1 determine the club activity area D and overlap area R using different rectangular areas B for each area in the frame image in this way, the sizes of the divided areas D and overlap area R in areas where the subject appears large become larger than the sizes of the divided areas D and overlap area R in areas where the subject appears small. As a result, even if the size of the subject differs depending on the position in the frame image, the subject can be detected appropriately at each position.

[0048] However, it is conceivable that the predetermined subject shown in the frame image at the time of initial setting may be smaller than the subject shown in the frame image during monitoring. In such a case, if the divided area D and the overlap area R are determined based on the small predetermined subject, the detection unit 132 may not be able to detect the subject in the overlap area R.

[0049] Therefore, the storage unit 12 may further store size information indicating the size of a predetermined subject. The size information stored in the storage unit 12 is information indicating the maximum size assumed as the size of the subject to be monitored.

[0050] In this case, the determination unit 133 corrects at least one of the divided region D and the overlap region R based on the size of the predetermined subject indicated by the size information. The determination unit 133 determines the correction amount of the overlap region R based on the size of the predetermined subject indicated by the size information. For example, the correction amount is determined based on the difference or ratio between the size of the predetermined subject indicated by the size information and the size of the subject to be monitored. The determination unit 133 corrects the divided region D and the overlap region R by adding or multiplying the determined correction amount to the overlap amount determined based on the width and height of the rectangular region B.

[0051] The size information may include, instead of the size of the predetermined subject, a correction value indicating the amount of correction for the overlap region R. In this case, the determination unit 133 determines the size of the overlap region R by adding or multiplying the determined overlap region R by the correction value.

[0052] The information processing device 1 corrects the rectangular area B corresponding to a specific subject based on the size of the subject to be monitored, and then determines the divided area D and overlap area R, thereby making it possible to set an appropriate divided area D taking into account the size of the subject being captured.

[0053] When an obstacle obscuring a monitored subject is included in a frame, even though the subject is not captured in the area containing the obstacle, performing a process to detect the subject in that area places an unnecessary load on the information processing device 1. The information processing device 1 can improve the efficiency of the detection process by setting an overlap area R excluding the area containing the obstacle.

[0054] Therefore, the detection unit 132 further detects the position of an obstacle, which is a subject included in a frame constituting the video data so as to obscure at least a portion of the predetermined subject. As an example, the storage unit 12 stores an obstacle detection model, which is a trained model trained to detect obstacles. The obstacle detection model is trained to output an area including an obstacle included in the image data based on training data that associates image data including an obstacle as a subject with information indicating the position of the obstacle. The detection unit 132 inputs frame images constituting the acquired video data into the obstacle detection model, and outputs an area including an obstacle included in the frame image.

[0055] The determination unit 133 determines the divided region D and the overlap region R based on the position of the obstacle detected by the detection unit 132. As an example, the determination unit 133 determines the divided region D so that the determined position of the obstacle becomes the starting point of the divided region D. In other words, the determination unit 133 determines the divided region so that the position of the detected obstacle becomes the edge of the divided region. Furthermore, the determination unit 133 determines the overlap region so that the position of the detected obstacle is not included in the overlap region.

[0056] The information processing device 1 may be configured to determine the overlap region R based on the speed at which a subject captured in moving image data moves.

[0057] The calculation unit 134 calculates the moving speed of a predetermined subject detected by the detection unit 132. The calculation unit 134 associates the same subject detected by the detection unit 132 in each of the frame images constituting the moving image data. The calculation unit 134 calculates the moving speed of the associated same subject. The calculation unit 134 may calculate the absolute moving speed as the moving speed, or may calculate the moving speed relative to the size of the area to be imaged by the imaging device 2.

[0058] The determination unit 133 determines the divided regions D and the overlapping regions R further based on the moving speed of the subject calculated by the calculation unit 134. The determination unit 133 determines the divided regions D and the overlapping regions R so that the greater the moving speed of the predetermined subject, the larger at least one of the divided regions D and the overlapping regions R becomes. When the moving speed of the predetermined subject is high, the determination unit 133 determines the divided regions D and the overlapping regions R so that the divided regions D and the overlapping regions R become larger compared to when the moving speed is low.

[0059] By configuring the information processing device 1 in this way, it is possible to divide the frame image so that the subject can be detected appropriately even when the subject is moving.

[0060] [Splitting the monitored video] After determining the division areas D and overlap areas R, the information processing device 1 divides the frame images constituting the moving image data based on the determined division areas D and overlap areas R, and monitors the moving image data. Hereinafter, the moving image data to be monitored will be referred to as target image data.

[0061] The storage unit 12 stores a surveillance model. The surveillance model is a trained model that is trained to detect a target subject, which is a subject to be monitored, included in the input image data, using image data having a lower resolution than the frame images included in the target video data. The surveillance model is trained to output an area in which the monitored subject included in the image data is included, based on training data that associates the image data with data indicating an area in which the monitored subject included in the image data is included. Note that the resolution of the image data that constitutes the training data is lower than the resolution of the frame images.

[0062] After the division area D and overlap area R are determined, the acquisition unit 131 acquires the target video data. The generation unit 135 generates a plurality of divided frame images based on the division area D and overlap area D determined by the determination unit 133 and the target video data. The divided frame images are image data obtained by dividing a frame image included in the target video data. The generation unit 135 generates divided frame images by dividing a frame image that makes up the video data into a plurality of pieces, based on division parameters stored in the storage unit 12.

[0063] The detection unit 132 detects target subjects included in the multiple divided frame images by inputting the multiple divided frame images into a surveillance model. Note that the target subject and the subject used as a reference for determining the divided region D and the overlap region R may be the same type of subject or different types of subjects. As an example, the information processing device 1 may be configured to determine the divided region D and the overlap region R based on a subject that is a person, and monitor a vehicle or the like as a target subject.

[0064] By configuring the information processing device 1 in this manner, the subject being monitored can be monitored based on an image divided so that a subject near a boundary can be detected in one of the areas, thereby making it possible to properly detect objects contained in video image data.

[0065] [Processing flow in information processing device 1] Fig. 7 is a flowchart showing the flow of processing in the information processing device 1. The flowchart shown in Fig. 7 shows processing when a divided area D and an overlap area R are determined while a person, who is a predetermined subject, moves through an area to be imaged by the imaging device 2. The flowchart shown in Fig. 7 starts from the point in time when the information processing device 1 receives an instruction to determine the divided area D. The instruction is input to the information processing device 1 by, for example, an installer of the imaging device 2.

[0066] The acquisition unit 131 acquires moving image data (S01). The detection unit 132 detects a predetermined subject in each frame image that constitutes the moving image data (S02). The determination unit 133 identifies the frame image in which the size of the rectangular area B is the largest (S03).

[0067] The determination unit 133 determines the divided area D and the overlap area R based on the size of the rectangular area B in the identified frame image (S04).

[0068] The determination unit 133 determines whether or not a termination condition is satisfied (S05). One example of the termination condition is that divided areas D have been set in all areas in the frame image. If the termination condition is satisfied (YES in S05), the information processing device 1 terminates the process.

[0069] If the termination condition is not satisfied (NO in S05), the determination unit 133 identifies a new frame image to be processed (S06). The frame image to be processed is, for example, a frame image whose undivided area includes rectangular area B. Once the information processing device 1 identifies a new frame image to be processed, the process proceeds to S04.

[0070] [Effects of information processing device 1] By configuring the information processing device 1 as described above, it is possible to divide an image so as to appropriately detect an object included in video data.

[0071] <Second embodiment> In the above embodiment, an example has been described in which the imaging device and the device for analyzing the captured image are configured as separate devices, but the processing described in the above embodiment as being performed in the information processing device 1 and the imaging device 2 may also be configured to be performed in the monitoring device 3.

[0072] 8 shows the configuration of a monitoring device 3 according to the second embodiment. The monitoring device 3 includes a communication unit 31, an imaging unit 32, a storage unit 33, and a control unit 34. The control unit 34 includes an acquisition unit 341, a detection unit 342, a determination unit 343, a calculation unit 344, and a generation unit 345.

[0073] The communication unit 31 is a communication interface for transmitting and receiving data to and from other devices via a network. The imaging unit 32 is a camera. The imaging unit 32 generates video image data and inputs it to the acquisition unit 341.

[0074] The storage unit 33 is a storage medium including a ROM, a RAM, an SSD, a hard disk drive, etc. The storage unit 33 stores in advance a program to be executed by the control unit 34. The control unit 34 is a processor such as a CPU or a GPU. The control unit 34 executes the program stored in the storage unit 33, thereby functioning as an acquisition unit 341, a detection unit 342, a determination unit 343, a calculation unit 344, and a generation unit 345.

[0075] The acquisition unit 341, the detection unit 342, the determination unit 343, the calculation unit 344, and the generation unit 345 perform the processes corresponding to the acquisition unit 131, the detection unit 132, the determination unit 133, the calculation unit 134, and the generation unit 135, respectively.

[0076] By configuring the monitoring device 3 in this way, it is possible to segment the image to be input to the object detection model so that an object can be appropriately detected with a simple configuration.

[0077] The present invention has been described above using embodiments, but the technical scope of the present invention is not limited to the scope described in the above embodiments, and various modifications and changes are possible within the scope of the gist of the present invention. For example, all or part of the device can be configured by functionally or physically distributing or integrating any unit. Furthermore, new embodiments resulting from any combination of multiple embodiments are also included in the embodiments of the present invention. The effects of the new embodiments resulting from the combination also have the effects of the original embodiments. [Explanation of symbols]

[0078] 1. Information processing equipment 2. Imaging device 3 Monitoring device 11 Communications Department 12 Storage section 13 Control Unit 131 Acquisition Department 132 Detector 133 Decision Section 134 Calculation Unit 135 Generation part 31 Communications Department 32 Imaging unit 33 Storage section 34 Control Unit 341 Acquisition Department 342 Detection unit 343 Decision Section 344 Calculation Unit 345 Generation part

Claims

1. an acquisition unit that acquires moving image data captured by the imaging device; a detection unit that detects a predetermined subject in a frame image included in the video data; a determination unit that determines a plurality of divided regions generated by dividing an image region of the frame image based on the size of the region in which the predetermined subject is detected, and an overlap region in which each of the plurality of divided regions overlaps with another adjacent divided region; An information processing device having the above.

2. the detection unit detects the predetermined subject included at different positions in the frame images, the determination unit determines the plurality of divided regions corresponding to the position where the predetermined subject is detected and the overlap region based on the position where the predetermined subject is detected and the size of the region where the predetermined subject is detected at the position. The information processing device according to claim 1 .

3. a storage unit for storing size information indicating the size of the predetermined subject; the determination unit corrects at least one of the divided regions and the overlap region based on the size of the predetermined subject indicated by the size information. The information processing device according to claim 1 or 2.

4. the detection unit further detects a position of an obstacle that is a subject included in a frame constituting the moving image data so as to obscure at least a part of the predetermined subject; the determination unit determines the divided regions and the overlapping region based on the position of the obstacle detected by the detection unit. The information processing device according to claim 1 or 2.

5. a calculation unit that calculates a moving speed of the predetermined subject detected by the detection unit, the determination unit determines the divided regions and the overlapping region so that at least one of the divided regions and the overlapping region becomes larger as the moving speed of the predetermined subject increases. The information processing device according to claim 1 or 2.

6. a generation unit that generates a plurality of divided frame images, which are image data obtained by dividing frames included in the target moving image data, based on the divided area and the overlap area determined by the determination unit and the target moving image data that is a moving image of a monitoring subject captured by the imaging device; the detection unit receives image data having a lower resolution than frames included in the target video data as input, and inputs the plurality of divided frame images into a trained model that has been trained to detect a target subject that is a subject to be monitored and included in the image data, thereby detecting the target subject included in the plurality of divided frames; 3. The information processing device according to claim 1 or 2.

7. An imaging unit; an acquisition unit that acquires moving image data captured by the imaging unit; a detection unit that detects a predetermined subject in a frame image included in the video data; a determination unit that determines a plurality of divided regions generated by dividing an image region of the frame image based on the size of the region in which the predetermined subject is detected, and an overlap region in which each of the plurality of divided regions overlaps with another adjacent divided region; A monitoring device having:

8. The computer executes acquiring moving image data captured by an imaging device; detecting a predetermined subject in a frame image included in the video data; determining a plurality of divided regions generated by dividing the image region of the frame image based on the size of the region in which the predetermined subject is detected, and an overlap region in which each of the plurality of divided regions overlaps with an adjacent divided region; An information processing method comprising:

Images