Information processing method, program, and information processing device

Abstract

This invention provides an information processing method that can generate an image from an image containing multiple people, with the main person in the composition being the central figure, according to the type of event depicted in the image. [Solution] An information processing method according to one embodiment of the present disclosure acquires an overall image in which multiple people are pictured, identifies the event type of the overall image, identifies a target person from the multiple people according to the identified event type, determines the area containing the target person as a cropping area, and generates individual images by cutting out the cropping area.

Description

Technical Field

[0001] This technology relates to an information processing method, a program, and an information processing apparatus.

Background Art

[0002] Conventionally, technologies for editing images have been proposed. For example, the trimming apparatus described in Patent Document 1 trims an image including a person's face to a panoramic size.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, the technology described in Patent Document 1 does not consider generating an image with the main person corresponding to the type of event shown in the image as the main subject of the composition from an image in which a plurality of people are shown.

[0005] The present disclosure has been made in view of such circumstances, and an object thereof is to provide an information processing method and the like capable of generating an image with the main person corresponding to the type of event shown in the image as the main subject of the composition from an image in which a plurality of people are shown.

Means for Solving the Problems

[0006] An information processing method according to an embodiment of the present disclosure acquires an overall image in which a plurality of people are shown, specifies the event type of the overall image, specifies a target person from the plurality of people according to the specified event type, determines a region including the target person as a trimming region, and generates an individual image obtained by cutting out the trimming region.

Effects of the Invention

[0007] In one embodiment of the information processing method of this disclosure, it is possible to generate an image from an image in which multiple people are depicted, with the main person in the composition being the main subject, according to the type of event depicted in the image. [Brief explanation of the drawing]

[0008] [Figure 1] This is a block diagram showing an example of the configuration of an information processing device. [Figure 2] This is an explanatory diagram showing an example of a complete image database. [Figure 3] This is an explanatory diagram showing an example of an individual image database. [Figure 4] This is an explanatory diagram showing an example of an individual image condition database. [Figure 5] This is an explanatory diagram illustrating an example of an event type identification model. [Figure 6] This is an explanatory diagram showing an example of a detection model. [Figure 7] This is an explanatory diagram showing an example of a target identification section. [Figure 8] This is an explanatory diagram showing an example of a trimming section. [Figure 9] This is an explanatory diagram showing an example of a super-resolution processing model. [Figure 10] This is an explanatory diagram illustrating an example of a segmentation model. [Figure 11] This is an explanatory diagram showing an example of a blurring processing unit and a synthesis unit. [Figure 12] This is an explanatory diagram showing an example of an individual image display screen. [Figure 13] This flowchart shows an example of individual image output processing. [Figure 14] This is an explanatory diagram showing an example of a detection model according to Embodiment 2. [Figure 15] This is an explanatory diagram showing an example of a target identification unit according to Embodiment 2. [Figure 16] This is an explanatory diagram showing an example of a trimming section according to Embodiment 2. [Figure 17] This is an explanatory diagram showing an example of a detection model according to Embodiment 3. [Figure 18] It is an explanatory diagram showing an example of the target person specifying unit according to Embodiment 3. [Figure 19] It is an explanatory diagram showing an example of the trimming unit according to Embodiment 3.

Mode for Carrying Out the Invention

[0009] (Embodiment 1) The information processing apparatus 1 according to Embodiment 1 specifies the event type of an overall image in which a plurality of persons are shown, specifies a target person from the plurality of persons according to the specified event type, determines a region including the target person as a trimming region, and generates an individual image obtained by cutting out the trimming region. Hereinafter, Embodiment 1 will be described with reference to the drawings.

[0010] FIG. 1 is a block diagram showing a configuration example of the information processing apparatus 1. The information processing apparatus 1 is, for example, a personal computer, and includes a processing unit 11, a storage unit 12, a communication unit 13, an input unit 14, and a display unit 15. Note that the information processing apparatus 1 may be a smartphone, a tablet terminal, a wearable device, or the like. The processing unit 11 is composed of a CPU (Central Processing Unit), MPU (Micro Processing Unit), GPU (Graphical Processing Unit), a quantum processor, or the like, and reads and executes a program P (program product) and a database stored in advance in the storage unit 12 to perform various control processes, arithmetic processes, and the like. Note that a database server or the like may be provided outside the information processing apparatus 1, and the database may be read from the database server or the like. Also, the function of the information processing apparatus 1 may be realized by a plurality of server devices or computers. Also, the information processing apparatus 1 may correspond to a node on a blockchain.

[0011] The storage unit 12 of the information processing apparatus 1 is, for example, a volatile memory and a non-volatile memory. The storage unit 12 stores a program P, a whole image DB121, an individual image DB122, an individual image condition DB123, an event type identification model M1, a detection model M2, a super-resolution processing model M3, and a segmentation model M4. Note that the program P may be provided to the information processing apparatus 1 using a storage medium 12a stored in a computer-readable manner. The storage medium 12a is, for example, a portable memory. Examples of the portable memory include a CD-ROM, a USB (Universal Serial Bus) memory, an SD card, a micro SD card, or a compact flash memory (registered trademark), etc. When the storage medium 12a is a portable memory, the processing element of the processing unit 11 may read the program P from the storage medium 12a using a reading device (not shown). The read program P is written into the storage unit 12. Further, the program P may be provided to the information processing apparatus 1 when the communication unit 13 communicates with an external device. Details of the whole image DB121, the individual image DB122, the event type identification model M1, the detection model M2, the super-resolution processing model M3, and the segmentation model M4 will be described later. The event type identification model M1, the detection model M2, the super-resolution processing model M3, or the segmentation model M4 may be stored in a server or a computer different from the information processing apparatus 1, and the information processing apparatus 1 may transmit an image to the server or the computer, and various processes may be performed in the server or the computer.

[0012] The communication unit 13 is a communication module or a communication interface for communicating with the imaging device 2 by wire or wirelessly, and is, for example, a wide-area wireless communication module such as LTE (registered trademark), 4G, or 5G. The processing unit 11 communicates with the imaging device 2 through the communication unit 13 via an external network N such as the Internet.

[0013] The input unit 14 receives input from the user, for example, an instruction to display an individual image. The input unit 14 may also receive input from the user, for example, an instruction to generate an individual image. The input unit 14 is composed of, for example, a keyboard or a mouse.

[0014] The display unit 15 displays the overall image acquired from the imaging device 2, as well as the generated individual images. The display unit 15 is composed of a display such as a liquid crystal panel or an organic EL (Electro-Luminescence) display.

[0015] The imaging device 2 is composed of, for example, a camera. The imaging device 2 takes a group photo including multiple people at an event such as a school event, track and field competition, or concert, and transmits the captured group photo to the information processing device 1. The imaging device 2 may store the captured group photo on a portable storage medium such as a USB memory, SD card, microSD card, or CompactFlash memory (registered trademark). In this case, the information processing device 1 can acquire the group photo by reading the portable storage medium. The information processing device 1 may also be equipped with an imaging unit, and the group photo may be taken by the imaging unit. If the information processing device 1 is a smartphone, tablet terminal, or wearable device, the group photo may be taken by the camera function of the information processing device 1. The type of event that can be captured in the group image is not limited to school events, track and field competitions, or concerts. The type of event may be a sports competition, a group tour, or a performance of dance or martial arts, etc.

[0016] Figure 2 is an explanatory diagram showing an example of the overall image DB 121. The overall image DB 121 stores information about the overall image. The management items (fields) of the overall image DB 121 include an overall image ID field, a shooting date and time field, an overall image field, and an event type field. The shooting date and time field stores the date and time when the overall image was taken. The overall image ID field stores a number (ID) assigned to distinguish the overall image. The overall image field stores the overall image acquired by the processing unit 11 from the shooting device 2, for example, in file format. The event type field stores the event type captured in the overall image, as identified by the event type identification model M1 (see Figure 5), which will be described later. Note that the event type field may also store an event type entered by the user. The overall image DB 121 may also include a shooting location field that stores the location where the overall image was taken.

[0017] Figure 3 is an explanatory diagram showing an example of the individual image DB 122. The individual image DB 122 stores information about individual images generated from the overall image. The management items (fields) of the individual image DB 122 include the individual image ID field, the individual image field, the subject ID field, and the overall image ID field. The individual image ID field stores a number (ID) assigned to distinguish individual images. The individual image field stores individual images (edited images) that have undergone various processing, for example, in file format. The subject ID field stores a number (ID) assigned to the person (subject) that is mainly depicted in the individual image. That is, the processing unit 11 associates the same subject ID with multiple individual images if individual images of the same subject are generated from multiple overall images based on the feature quantities of the subject detected by the detection model M2 (see Figure 6), which will be described later. The overall image ID field stores the overall image ID of the overall image from which the individual images were generated.

[0018] Figure 4 is an explanatory diagram showing an example of the Individual Image Condition DB123. The Individual Image Condition DB123 stores conditions for subjects, non-subjects, and cropping areas according to the event type. The management items (fields) of the Individual Image Condition DB123 include, for example, an event type field, a subject field, a non-subject field, and a cropping area field. The event type field stores the event type. The subject field stores the conditions for the person who will be the subject according to the event type. The non-subject field stores the conditions for the person who will be the non-subject, according to the event type. The cropping area field stores the conditions for the cropping area according to the event type. Note that the conditions for subjects, non-subjects, and cropping areas according to the event type may also be incorporated into program P.

[0019] For example, if the event type is "Activities at a school, nursery school, or kindergarten (school, etc.)", the target audience is "children", the non-target audience is "adults", and the cropping area is "an area cropped so that the entire body of the child is included, with the child at the center." If the event type is track and field, the target audience is "people wearing bibs", the non-target audience is "people without bibs", and the cropping area is "an area cropped so that the area on the side of the subject in the direction of movement is wider than the area on the side not in the direction of movement." If the event type is a concert, the target audience is "people playing musical instruments", the non-target audience is "people not playing musical instruments", and the cropping area is "an area cropped so that both the target audience and the instrument they are playing are included."

[0020] Figure 5 is an explanatory diagram showing an example of the event type identification model M1. The event type identification model M1 is composed of an object detection machine learning model that includes a neural network such as CNN (Convolutional Neural Network), R-CNN (Regions-CNN), Fast R-CNN, Faster R-CNN, SSD (Single Shot Multibook Detector), or YOLO (You Only Look Once). The event type identification model M1 is trained to output the event type when a whole image is input, using training data that associates images with their respective event types.

[0021] When the event type identification model M1 is composed of a model including a neural network such as a CNN, the event type identification model M1 has multiple neurons that accept pixel values ​​of the overall image as input and passes the input pixel values ​​to the hidden layer. The hidden layer has multiple neurons that extract image features from the overall image and passes the extracted image features to the output layer. The output layer outputs the event type of the overall image based on the image features. In this embodiment, the overall image contains multiple children, and the event type identification model M1 outputs "Activities at school, nursery school, or kindergarten (school, etc.)" as the event type of the overall image. The event type identification model M1 may also output more subdivided events such as sports day, field trip, or cultural festival as event types. Furthermore, the processing unit 11 of the information processing device 1 may identify the event type using an image embedding vector obtained by inputting the overall image into the autoencoder of a VLM (Vision Language Model).

[0022] Figure 6 is an explanatory diagram showing an example of detection model M2. The processing unit 11 detects people in the overall image by inputting the overall image to detection model M2. The processing unit 11 also identifies subjects and non-subjects using the subject identification unit 111. Detection model M2 is composed of an object detection machine learning model that includes a neural network such as CNN, R-CNN, Fast R-CNN, Faster R-CNN, SSD, or YOLO. Detection model M2 uses training data that associates images with regions containing people and attributes of people in the image, and outputs regions containing people and attributes of people when an overall image is input. Note that detection model M2 may be composed of a model integrated with event type identification model M1. Detection model M2 outputs attributes based on the category of person or their action, such as child, adult, athlete, non-athlete, performer, or non-performer.

[0023] When the detection model M2 is composed of a neural network such as a CNN, the detection model M2 has multiple neurons that accept pixel values ​​as input to the entire image and passes the input pixel values ​​to the hidden layer. The hidden layer has multiple neurons that extract image features from the entire image and passes the extracted image features to the output layer. Based on the image features, the output layer outputs the area in the entire image where a person is present. In the image showing the output of the detection model M2 in Figure 6, the area containing the person is indicated by a solid bounding box.

[0024] Figure 7 is an explanatory diagram showing an example of the subject identification unit 111. The subject identification unit 111, as a functional unit of the processing unit 11, executes the processing described later. The processing unit 11 inputs the entire image, in which the area containing people has been output, to the subject identification unit 111. The subject identification unit 111 refers to the individual image condition DB 123 according to the event type output by the event type identification model M1 and identifies non-target individuals based on the attributes of the people (identification based on attributes). As shown in Figure 7, if the event type is an activity at a school, etc., the subject identification unit 111 identifies adults among the detected people as non-target individuals. In the overall image after identification based on attributes, the area containing non-target individuals is indicated by a dashed bounding box. Thereafter, the bounding boxes of people identified as non-target individuals are indicated by dashed lines, and the bounding boxes of people identified as target individuals are indicated by solid lines.

[0025] The subject identification unit 111 identifies individuals who are overlapped by other people or objects in the foreground as non-subjects. The subject identification unit 111 may also identify individuals whose faces are overlapped by other people or objects as non-subjects. Furthermore, the subject identification unit 111 identifies individuals whose bounding boxes touch the edges of the overall image, individuals who are out of focus, and individuals whose bounding boxes are less than a predetermined size (for example, 2% of the overall image size) as non-subjects (excluded from subjects) (identification based on how they appear in the image). The subject identification unit 111 (processing unit 11) identifies non-subjects by evaluating the size, clarity, or bounding box size of the portion of the child's body that is visible in the overall image. Among the individuals visible in the overall image, those who were not identified as non-subjects by the above processing are identified as subjects.

[0026] Figure 8 is an explanatory diagram showing an example of the trimming unit 112. The trimming unit 112, as a functional unit of the processing unit 11, performs the processing described later. The trimming unit 112 refers to the individual image condition DB 123 according to the event type output by the event type identification model M1 and determines the trimming area to be an individual image. As shown in Figure 8, if the event type is an activity at school, the trimming area is determined to be an area that is cropped so that the entire body of the child is included, with the child at the center. The conditions for the trimming area may also be determined by the ratio of the surrounding area (extra area) to the bounding box in the individual image, or the ratio or number of people other than the subject in the individual image. If multiple people among those in the overall image are identified as subjects, the trimming area is cut out and individual images for each subject are generated. The trimming unit 112 assigns an individual image ID to each generated (output) individual image.

[0027] Figure 9 is an explanatory diagram showing an example of a super-resolution processing model M3, etc. The processing unit 11 inputs individual images to the resizing unit 113 and reduces the image size of the individual images (resizing process). Resizing can speed up the processing described later.

[0028] The processing unit 11 inputs the resized individual images to the super-resolution processing model M3. The super-resolution processing model M3 is composed of a machine learning model that performs image generation, such as a GAN (Generative Adversarial Network). The super-resolution processing model may also be composed of a machine learning model such as SRCNN (Super-Resolution CNN), Pix2pix, or CUT (Contrastive Learning for Unpaired Image-to-Image Translation), or a generative model such as VLM. When the super-resolution processing model M3 is composed of a GAN, the super-resolution processing model M3 includes a generator and a classifier. The generator of the super-resolution processing model M3 generates a high-resolution image based on the features of the individual images. The classifier compares the image generated by the generator with the input individual images and determines whether the generated image is genuine or fake compared to the individual images. The generator obtains the classifier's judgment result and performs image regeneration based on the obtained judgment result. The generator generates images until the judgment result output by the classifier becomes "true," and the super-resolution processing model M3 outputs the images that are judged as "true" as super-resolution images.

[0029] Figure 10 is an explanatory diagram showing an example of a segmentation model M4. The processing unit 11 inputs the super-resolution image output by the super-resolution processing model M3 to the segmentation model M4. The segmentation model M4 is composed of a machine learning model with semantic segmentation capabilities, such as FCN (Fully Convolutional Network), seg-net, or FPN (Feature Pyramid Networks). The segmentation model M4 uses training data that associates images with labels indicating whether or not each pixel in the image represents a person, and when an individual image (super-resolution image) is input, it outputs whether or not each pixel represents a person, that is, the region of pixels that represent a person in the individual image (person region).

[0030] When the segmentation model M4 is composed of a neural network such as an FCN, the segmentation model M4 has multiple neurons that accept pixel values ​​from the overall image as input, and passes the input pixel values ​​to the hidden layer. The hidden layer has multiple neurons that extract image features from the overall image, and passes the extracted image features to the output layer. The output layer outputs the semantic range of the subject based on the image features. In the example shown in Figure 10, the segmentation model M4 cuts out the person region of the subject from the individual images and outputs it.

[0031] Figure 11 is an explanatory diagram showing an example of a blurring processing unit 114 and a compositing unit 115. The blurring processing unit 114 and the compositing unit 115 function as functional units of the processing unit 11 and perform the processing described later. The processing unit 11 receives individual images that have not undergone super-resolution processing as input to the blurring processing unit 114. The blurring processing unit 114 outputs a blurred image in which the individual images have been blurred (blurred).

[0032] The synthesis unit 115 of the processing unit 11 acquires the blurred image output by the blurring processing unit 114 and the person region output by the segmentation model M4, and outputs an edited image in which the person region is superimposed on the blurred image. The person region is superimposed on the blurred image by matching the position information of each pixel in the individual image contained in the person region with the position of each pixel in the blurred image. As a result, the processing unit 11 outputs an edited image in which the background region other than the person region in the individual image has been blurred.

[0033] Figure 12 is an explanatory diagram showing an example of an individual image display screen. The processing unit 11 displays the individual images (edited images) that have undergone the various processing steps described above on the display unit 15. In addition, the processing unit 11 displays the overall image on which the individual images were generated, along with the individual images, on the display unit 15. The processing unit 11 may also display images at each stage of processing for the overall image or individual images as shown in Figures 5 to 11 on the display unit 15. Furthermore, the processing unit 11 may accept input instructions for outputting individual images to other terminals or to a printing device on the individual image display screen.

[0034] Figure 13 is a flowchart illustrating an example of individual image output processing. The processing unit 11 of the information processing device 1 reads the overall image from the overall image DB 121 (S1). The processing unit 11 inputs the overall image to the event type identification model M1 (S2) and outputs the event type (S3). The processing unit 11 inputs the overall image to the detection model M2 (S4) and outputs the region containing a person and the attributes of the person in the overall image (S5).

[0035] The processing unit 11 of the information processing device 1 reads the conditions for subjects, non-subjects, and trimming regions from the individual image condition DB 123 according to the event type output by the event type identification model M1 (S6). Based on the attributes of the person output by the detection model M2, the processing unit 11 excludes persons who meet the conditions for non-subjects from the subjects (they are designated as non-subjects) (S7). The processing unit 11 excludes persons from the subjects who are overlapping with other people or objects in the foreground of the overall image (S8). The processing unit 11 excludes persons whose bounding box touches the edge of the overall image, persons who are out of focus, and persons whose bounding box is less than a predetermined size (for example, 2% of the size of the overall image) from the subjects (S9). The processing unit 11 determines the trimming region according to the conditions for the trimming region based on the event type of the overall image (S10). The processing unit 11 generates individual images with the trimmed region extracted (S11).

[0036] The processing unit 11 performs resizing on individual images (S12). The processing unit 11 inputs the individual images into the super-resolution processing model M3 (S13) and outputs a super-resolution image (S14). The processing unit 11 inputs the super-resolution image into the segmentation model M4 and extracts and outputs the person region of the subject in the super-resolution image (S15).

[0037] The processing unit 11 performs a blurring process on the individual images (S16). The processing unit 11 outputs an edited image by superimposing the person region extracted from the super-resolution image onto the blurred individual images (S17).

[0038] The processing unit 11 stores the output edited image in the individual image DB 122 (S18). The processing unit 11 then displays the edited image on the display unit 15 (S19) and terminates the process.

[0039] According to the configuration and processing of this embodiment, it is possible to identify the main person in an image corresponding to the type of event depicted in the image, and generate an image in which each person is the main subject of the composition. Furthermore, it is possible to generate individual images in which the subjects stand out more through super-resolution processing and blurring processing.

[0040] (Embodiment 2) Embodiment 2 describes the identification of subjects and generation of individual images when the event type of the overall image is track and field. As shown in the individual image conditions DB123 (see Figure 4), when the event type of the overall image is track and field, subjects are "people wearing bibs," non-subjects are "people without bibs," and the cropping area is "an area cropped so that the area on the direction of movement of the subject is wider than the area on the non-direction of movement."

[0041] Figure 14 is an explanatory diagram showing an example of the detection model M2 according to Embodiment 2. The detection model M2 according to Embodiment 2 is trained to output (detect) the region containing a person and the location of the bib when an overall image is input, using training data that associates an image with the location of the bib in the image. The detection model M2 may also detect uniforms, hats, gloves, shoes, or other athletic equipment related to the competition.

[0042] Figure 15 is an explanatory diagram showing an example of the subject identification unit 111 according to Embodiment 2. The subject identification unit 111 according to Embodiment 2 identifies individuals who are not wearing bibs among the individuals detected by the detection model M2 as non-subjects. The subject identification unit 111 may also identify individuals who are not wearing competition uniforms, hats, gloves, or shoes, or who are not carrying competition equipment, as non-subjects. Furthermore, similar to Embodiment 1, the subject identification unit 111 identifies individuals who are overlapping with other people or objects in the foreground, individuals whose bounding boxes touch the edges of the overall image, individuals who are out of focus, and individuals whose bounding boxes are less than a predetermined size (for example, 2% of the overall image size) as non-subjects.

[0043] Figure 16 is an explanatory diagram showing an example of a trimming unit 112 according to Embodiment 2. The trimming unit 112 according to Embodiment 2 refers to an individual image condition DB and determines the trimming area to be a region that is cut out so that the area on the side of the subject's direction of movement is wider than the area on the side of the non-direction of movement. By making the trimming area a region that is cut out so that the area on the side of the subject's direction of movement is wider than the area on the side of the non-direction of movement, the sense of dynamism of the subject in the individual image is highlighted. In the example shown in Figure 16, the processing unit 11 identifies the direction of the subject's movement to the left, for example, based on the orientation of the subject's face. The trimming unit 112 determines the trimming area so that the area Rf on the left side (the side of the subject's direction of movement) is wider than the area Rb on the right side (the side of the subject's non-direction of movement) relative to the bounding box.

[0044] According to the processing method of this embodiment, it is possible to generate individual images of athletes from an overall image of a track and field event.

[0045] (Embodiment 3) Embodiment 3 describes the identification of subjects and the generation of individual images when the event type of the overall image is a concert. As shown in the individual image condition DB123 (see Figure 4), when the event type of the overall image is a concert, subjects are "people playing musical instruments," non-subjects are "people not playing musical instruments," and the cropping area is "an area cropped to include the subjects and the instruments they play."

[0046] Figure 17 is an explanatory diagram showing an example of the detection model M2 according to Embodiment 3. The detection model M2 according to Embodiment 3 is trained to output (detect) the region containing a person and the position of the instrument when an overall image is input, using training data that associates an image with the position of the instrument in the image. The detection model M2 may also detect the performer's clothing or tie, etc.

[0047] Figure 18 is an explanatory diagram showing an example of the subject identification unit 111 according to Embodiment 3. The subject identification unit 111 according to Embodiment 3 identifies individuals who are not playing a musical instrument as non-subjects among the people detected by the detection model M2. The subject identification unit 111 may also identify individuals who are not wearing the performer's costume or tie as non-subjects. Furthermore, similar to Embodiment 1, the subject identification unit 111 identifies individuals who are overlapping with other people or objects in the foreground, individuals whose bounding box touches the edge of the overall image, individuals who are out of focus, and individuals whose bounding box is less than a predetermined size (for example, 2% of the overall image size) as non-subjects.

[0048] Figure 19 is an explanatory diagram showing an example of a trimming unit 112 according to Embodiment 3. The trimming unit 112 according to Embodiment 3 refers to an individual image condition DB and determines the trimming area to be a region that includes the subject and the instrument that the subject is playing. By making the trimming area a region that includes the subject and the instrument that the subject is playing, it is possible to generate individual images that show the subject playing.

[0049] According to the processing of this embodiment, it is possible to generate individual images of performers from an overall image of a concert.

[0050] The embodiments disclosed herein should be considered in all respects as illustrative and not restrictive. The technical features described in each embodiment can be combined with each other, and the scope of the present invention is intended to include all modifications within the claims and scope equivalent to the claims. Furthermore, the independent and dependent claims described in the claims can be combined with each other in any combination, regardless of the form of reference. In addition, the claims use a multi-claim format in which claims refer to two or more other claims (multi-claim format), but are not limited to this. They may also be described using a multi-claim format in which at least one multi-claim refers to another multi-claim (multi-multi-claim format). [Explanation of symbols]

[0051] 1: Information Processing Device 11: Processing Section 111: Target Identification Department 112: Trimming section 113: Resizing section 114: Blur processing 115: Synthesis part 12: Storage section 12a:Storage medium 13: Communications Department 14: Input section 15: Display section 121: Overall Image Database 122: Individual Image Database 123: Individual Image Conditions Database M1: Event type identification model M2: Detection Model M3: Super-resolution processing model M4: Segmentation Model 2: Imaging device N: Network P: Program

Claims

1. Obtain a full image containing multiple people, Identify the event type of the aforementioned overall image Depending on the identified event type, the target person is selected from the aforementioned multiple individuals. The area including the aforementioned subject is determined to be the trimming area. Generate individual images by cutting out the aforementioned trimmed area. Information processing methods.

2. If the aforementioned event type is an activity at a school, nursery school, or kindergarten, adults will be identified as non-target individuals who are not among the aforementioned target groups. The information processing method according to claim 1.

3. If the event type is track and field, individuals without bib numbers are identified as non-target individuals. The information processing method according to claim 1.

4. Identify individuals who are overlapped by other people or objects as non-target individuals. The information processing method according to any one of claims 1 to 3.

5. In the aforementioned overall image, individuals whose size is less than a predetermined size are identified as non-target individuals. The information processing method according to any one of claims 1 to 3.

6. The trimming area is determined according to the type of event. The information processing method according to any one of claims 1 to 3.

7. Super-resolution processing is performed on the aforementioned individual images using a machine learning model. The information processing method according to claim 1.

8. After performing a resizing process on the individual images, super-resolution processing is performed. The information processing method according to claim 7.

9. The output is an edited image in which the background area other than the area containing the subject in the individual image has been blurred. The information processing method according to claim 7 or 8.

10. The person region is extracted from the individual images that have undergone super-resolution processing. A blurring process is performed on the aforementioned individual images. The extracted area of ​​the person is superimposed onto the blurred individual image. By doing so, the edited image is output. The information processing method according to claim 9.

11. Obtain a full image containing multiple people, Identify the event type of the aforementioned overall image Depending on the identified event type, the target person is selected from the aforementioned multiple individuals. The area including the aforementioned subject is determined to be the trimming area. Generate individual images by cutting out the aforementioned trimmed area. A program that instructs a computer to perform a process.

12. Obtain a full image containing multiple people, Identify the event type of the aforementioned overall image Depending on the identified event type, the target person is selected from the aforementioned multiple individuals. The area including the aforementioned subject is determined to be the trimming area. Generate individual images by cutting out the aforementioned trimmed area. Processing unit that executes the process An information processing device equipped with the following features.

Images