JP7542165B1 - Information processing device, information processing method, and information processing program - Google Patents

Abstract

[Problem] To provide an information processing device, information processing method, and information processing program capable of estimating the area of a target object to be annotated. [Solution] The information processing device includes an acquisition unit that acquires an image, a reception unit that receives instructions about the target object, an estimation unit that estimates the area of the target object recorded in the image acquired by the acquisition unit based on the instructions received by the reception unit, and an output control unit that controls the area of the target object estimated by the estimation unit to be output. [Selected Figure] Figure 1

Description

Applicable under Article 30, Paragraph 2 of the Patent Act Publisher: SoftBank Corp. Published at: https://www.docs.tsk-pf.com/annotation-screen/ai-annotation/prompt-annotation Published on: July 5, 2023

Patent Law Article 30, Paragraph 2 applied Distributor SoftBank Corp. Distributed at Tokyo Big Sight Distributed on July 11, 2023 1st AI World Summer

Patent Law Article 30, Paragraph 2 applies. Distributor: SoftBank Corp. Disclosure location: http://ans. bb. local/#/detail/d21c925b80ffcddccf9818953e72988d? keyword=tasuki%20annotation Disclosure date: September 5, 2023

Patent Law Article 30, Paragraph 2 applicable Publisher SoftBank Corp. Publication location https://tsk-pf. com/annotation-tool Publication date November 2, 2023

Patent Law Article 30, Paragraph 2 applies. Publisher: iSmiley Co., Ltd. Published at: https://aiSmiley.co.jp/product/sbi_tasuki-annotation-tool/ Published on: January 26, 2024

Applicable under Article 30, Paragraph 2 of the Patent Act. Disclosed by SoftBank Corp. Published at https://biz.tm.softbank.jp/pg12450-web-doc-entry-tasuki.html Published on January 30, 2024

Applicable under Article 30, Paragraph 2 of the Patent Act. Disclosed by SoftBank Corp. Disclosed at https://www.softbank.jp/biz/news/cloud/20240130/ Disclosed on January 30, 2024

The present disclosure relates to an information processing device, an information processing method, and an information processing program.

Conventionally, there is a device that supports image annotation. The device classifies multiple target areas of a target image that are candidates for annotation based on the features that appear in the target image, and generates classification information. The device visualizes the classification information and displays it so that it can be compared with the target image.

JP 2022-131937 A

However, the task of annotating is time-consuming, and there is a need to automatically identify targets to which annotations should be added.

The present disclosure provides an information processing device, an information processing method, and an information processing program that can estimate the area of a target object to be annotated.

An information processing device of one embodiment includes an acquisition unit that acquires an image, a reception unit that receives instructions about a target object, an estimation unit that estimates the area of the target object recorded in the image acquired by the acquisition unit based on the instructions received by the reception unit, and an output control unit that controls the area of the target object estimated by the estimation unit to be output.

The information processing device, information processing method, and information processing program disclosed herein can estimate the area of a target object to which annotation is to be added.

11 is a diagram for explaining a screen displayed based on control of an information processing device according to an embodiment. FIG. FIG. 1 is a block diagram illustrating an information processing device according to an embodiment. FIG. 2 is a diagram for explaining an example of an image (image information). 4 is a diagram for explaining an example of an instruction content (prompt) regarding the image (image information) illustrated in FIG. 3 . FIG. 1 is a flowchart illustrating an information processing method according to an embodiment.

One embodiment is described below.

[Overview of information processing device 100]
First, an overview of an information processing device 100 according to an embodiment will be described.
FIG. 1 is a diagram for explaining a screen displayed under the control of an information processing device 100 according to an embodiment.

The information processing device 100 may be configured as an estimation device (identification device) that estimates (identifies) an area of an object (target object) that is recorded in the image 200 and corresponds to a character string based on, for example, an instruction content 141 (prompt) that is a received character string. The information processing device 100 may also be configured as an annotation device that adds an annotation to the estimated area of the object (target object). The information processing device 100 is not limited to the device of the above example, and may constitute various devices.
The information processing device 100 may be a computer such as a server, a desktop, a laptop, a tablet, or a smartphone.

The information processing device 100 acquires an image 200. In the image 200, one or more objects are recorded.
1 , the information processing device 100 acquires an image 200 recording a road and a vehicle (e.g., objects such as a passenger car 201 and a truck 202) traveling on the road, and displays it on the display unit 133. The information processing device 100 also displays an interface (e.g., an input interface 140 using chat) for inputting instructions (prompts) to the information processing device 100 on the display unit 133.

The information processing device 100 receives, for example, a prompt 141 in the form of a character string (text) or the like via the input interface 140. The character string may have various contents, and may also be the contents of an object (target object) whose area is to be estimated. That is, as an example, when a passenger car 201 and a truck 202 are recorded as objects in an image 200, and the area of the passenger car 201 in the image 200 is to be estimated, the information processing device 100 receives "passenger car" (prompt) 141 as the character string.

The information processing device 100 estimates the area of the target object recorded in the image 200 based on the instruction content 141 received as described above. The information processing device 100 estimates the area of the object (target object) corresponding to the instruction content 141 recorded in the image 200 by using various processes including, for example, a known object recognition process. As an example, when there are a regular car 201 and a truck 202 as objects recorded in the image 200 and the area of the regular car 201 in the image 200 is to be estimated, the information processing device 100 estimates (specifies) the area of the regular car 201 (target object) in the image 200 when it receives the character string "regular car" (instruction content) 141.
In this case, the information processing device 100 may identify a region of the target object having a size equal to or larger than a threshold value based on the size of the estimated (identified) region of the target object and a threshold value input via the input interface 140.

The information processing device 100 outputs the area of the target object estimated (identified) as described above. For example, the information processing device 100 displays the area of the target object in the image 200 in a color different from other areas. In the example shown in FIG. 1, the area of the target object (a passenger car 201) in the image 200 is shaded and other areas (e.g., truck 202, etc.) are not shaded, so that the area of the target object (a passenger car 201) is shown in a different manner from other areas (e.g., truck 202, etc.).

Furthermore, the information processing device 100 annotates the region of the target object identified in the image 200 based on the instruction content 141 (prompt) of the character string received as described above. That is, in the example shown in Fig. 1, the information processing device 100 may identify the region of the target object (standard-sized car 201) marked with diagonal lines based on "standard-sized car" input as the character string (instruction content 141 (prompt)), and may annotate the region of the target object (standard-sized car 201) with "standard-sized car."
Note that the objects to which annotations are added are not limited to the passenger car 201 and the truck 202, but may be various other objects.

[Details of information processing device 100]
Next, the information processing device 100 according to an embodiment will be described in detail.
FIG. 2 is a block diagram for explaining the information processing device 100 according to an embodiment.
FIG. 3 is a diagram for explaining an example of an image 200 (image information).
FIG. 4 is a diagram for explaining an example of an instruction content (prompt) regarding the image 200 (image information) illustrated in FIG.

The information processing device 100 includes, for example, an input unit 121, a communication unit 131, a storage unit 132, a display unit 133, and a control unit 110. The communication unit 131, the storage unit 132, and the display unit 133 may be an embodiment of the output unit. The control unit 110 includes, for example, an acquisition unit 111, a reception unit 112, an estimation unit 113, an output control unit 114, a learning unit 115, and an AI unit 116. The control unit 110 may be configured, for example, by a calculation processing unit of the information processing device 100. The control unit 110 (for example, a calculation processing unit, etc.) may realize the functions of each unit (for example, the acquisition unit 111, the reception unit 112, the estimation unit 113, the output control unit 114, the learning unit 115, and the AI unit 116, etc.) by, for example, appropriately reading and executing various programs stored in the storage unit 132, etc. That is, the functions of each unit may be realized by computer implementation.

The input unit 121 may be, for example, a graphical user interface (GUI) or the like. The input unit 121 may also be, for example, a keyboard and a mouse or the like.

The communication unit 131 is, for example, a communication interface capable of transmitting and receiving various information to and from a device (external device) (not shown) external to the information processing device 100.

The storage unit 132 may store, for example, various information and programs. Examples of the storage unit 132 may be a memory, a solid state drive, a hard disk drive, etc. The storage unit 132 may be, for example, a storage area and a server on the cloud, etc.

The display unit 133 is, for example, a display capable of displaying various characters, symbols, images, etc.

The acquisition unit 111 acquires an image 200 (image information).
The acquisition unit 111 acquires, for example, an image 200 (image information) stored in the storage unit 132 .
Furthermore, the acquisition unit 111 acquires the image 200 (image information) from an external device (not shown), for example, via the communication unit 131. The external device may be, for example, a server or a user terminal. The user terminal is a terminal used by a user of the information processing device 100, and may be a desktop, a laptop, a tablet, a smartphone, or the like.
In addition, the acquisition unit 111 may acquire the image 200 (image information) from an external memory (not shown) in which the image 200 (image information) is stored, for example, when the external memory is connected to an interface (not shown) of the information processing device 100.
Image 200 may be a still image or a video image. Image 200 captures one or more objects.
As an example, as shown in FIG. 3, the acquisition unit 111 may acquire an image 200 (image information) recording a road and a vehicle traveling on the road (e.g., objects such as a passenger car 201 and a truck 202).

The reception unit 112 receives, for example, instruction contents 141 (prompt) regarding the target object via the input interface 140. The reception unit 112 may receive, for example, instruction contents 141 (prompt) in the form of a character string (text) regarding the target object. The character string (text) may be of various kinds, and may be, for example, the content of an object (target object) whose area is to be estimated. The input interface 140 may be, for example, an interface that receives instruction contents 141 (prompt) using chat.

That is, as an example, as shown in FIG. 4, objects recorded in an image 200 include a passenger car 201 and a truck 202, and when estimating the area of the passenger car 201 in the image 200, the information processing device 100 accepts a character string "Passengers" (instruction content) 141. In response to "Passengers" being input to the input interface 140, the information processing device 100 may display content 142 such as "I will infer" on the display unit 133 (input interface 140) as a chat response for estimating the area of the passenger car 201.

Similarly, as an example, objects recorded in image 200 include a passenger car 201 and a truck 202, and when estimating the area of truck 202 in image 200, information processing device 100 accepts "truck" (instruction content) as the character string.
Similarly, as an example, when objects recorded in image 200 include a passenger car 201 and a truck 202, and the area of both the passenger car 201 and the area of the truck 202 in image 200 are to be estimated, the information processing device 100 may accept the character strings "passenger car" and "truck" (multiple instruction contents).

The estimation unit 113 estimates the area of the target object recorded in the image 200 acquired by the acquisition unit 111, based on the instruction content 141 accepted by the acceptance unit 112. The estimation unit 113 estimates the area of the object (target object) corresponding to the instruction content 141, recorded in the image 200, by using various processes including, for example, a known object recognition process.
When using the example described above, that is, when the objects recorded in image 200 are a passenger car 201 and a truck 202, and the area of passenger car 201 in image 200 is to be estimated, when estimation unit 113 receives a character string "passenger car" (instruction content) 141 (see Figure 4), it estimates (identifies) the area of passenger car 201 (target object) in image 200.
Similarly, as an example, when objects recorded in image 200 include a passenger car 201 and a truck 202 and the area of truck 202 in image 200 is to be estimated, when the estimation unit 113 receives the character string "truck" (instruction content), it estimates (identifies) the area of truck 202 (target object) in image 200.
Similarly, as an example, when objects recorded in image 200 include a passenger car 201 and a truck 202, and the area of both the passenger car 201 and the area of the truck 202 in image 200 are to be estimated, the information processing device 100 may, when receiving the character strings "passenger car" and "truck" (multiple instruction contents), estimate (identify) the areas of the passenger car 201 and the truck 202 (multiple target objects) in image 200.
When the estimation is successful, the estimation unit 113 may display a chat reply with content 144 such as "Successful" on the input interface 140. When the estimation is unsuccessful, the estimation unit 113 may display a chat reply with content such as "Please re-enter the prompt," for example.

The estimation unit 113 may estimate at least one of the pixel region of the target object in the image 200 and the contour boundary of the target object as the region of the target object. That is, the estimation unit 113 may estimate at least one of a group of pixel regions and contour boundaries in the target object. The pixel region of the target object may be a surface region that is inside the contour of the target object. The contour boundary of the target object can be rephrased as the contour of the target object.

The estimation unit 113 may assign a value according to the size of the area of the target object. For example, the estimation unit 113 may assign a value of 100 to the largest area of the areas of the multiple target objects estimated as described above, assign a value of 0 to the smallest area, and assign a value greater than 0 and less than 100 to each area from the smallest area to the largest area according to the size of the area. Note that the values (range of values) are not limited to the range from 0 to 100 described above, and various values (ranges of values) may be assigned.

As an example of the specific processing contents of the estimation unit 113 described above, the estimation unit 113 may first detect a rectangular area of each object (each target object) in the image 200 corresponding to the instruction contents 141 (prompt) to generate a rectangular area image, and then perform contour approximation for the object in each rectangular area image to obtain the coordinates of the contour. In this case, the estimation unit 113 may generate a rectangular area image by, for example, processing such as AI. As an example of processing such as AI, when a "reliability score" that specifies the reliability of the object (target object) is input, the estimation unit 113 may specify an object (target object) that is estimated according to the reliability score, and generate an image of a rectangular area around the specified object (target object). As an example, the reliability score may be a value in the range from 0 to 1. Furthermore, the estimation unit 113 may surround the object (object contour) with a polygon (polygon) based on the acquired coordinates, and obtain the coordinate points of the polygon. The estimation unit 113 may acquire the coordinate points of the polygon as described above for each of the multiple objects. That is, the estimation unit 113 may estimate at least one of the contour (contour region) of the object and the inside of the object (pixel region of the object) that is inside the contour (contour region) by using the coordinate points of the polygon.
Next, the estimation unit 113 scores the size of each of the polygons (areas enclosed by the coordinate points of the polygons) of the multiple objects based on the size of the largest polygon and the size of the smallest polygon. That is, as an example, the estimation unit 113 assigns a value between 0 and 100 depending on the size of each of the multiple polygons, for example, assigning a value of 100 to the largest polygon and a value of 0 to the smallest polygon.

When the receiving unit 112 receives the instruction content 141, the estimation unit 113 may generate a different thread (i.e., a page, etc.) for each instruction content 141 (for each type of target object) and generate an area for one target object in each thread (each page).

The output control unit 114 controls the output unit to output the area of the target object estimated by the estimation unit 113. Here, the output unit may be, for example, the communication unit 131, the memory unit 132, the display unit 133, etc.

As an example, when the estimation unit 113 estimates the areas of multiple target objects (e.g., passenger car 201 (or truck 202)) in the image 200, the output control unit 114 may control the output unit to output in a manner that identifies the area of the estimated passenger car 201 (or truck 202) (target object) in the image 200.

As another example, the output control unit 114 may control the output unit to, for example, when the estimation unit 113 estimates the areas of each of a plurality of target objects (e.g., a passenger car 201 and a truck 202) in the image 200, output the area of the target object corresponding to the input character string (either "a passenger car" or "a truck") via the input interface 140, i.e., output the area of the passenger car 201 (or the truck 202) (target object) in the image 200 when "a passenger car" (or "a truck") is input as a character string for identifying the target object.

As another example, the output control unit 114 may, for example, when the estimation unit 113 estimates the area of a target object (e.g., a passenger car 201) in the image 200, and then the estimation unit 113 estimates the area of a target object (e.g., a truck 202) in the same image 200, create a thread (e.g., a screen of the passenger car 201 and a screen of the truck 202) for each estimated target object (each of the passenger car 201 and the truck 202), and control the output unit to output one (one) or both (multiple) of the multiple threads. In this case, the output control unit 114 may, for example, control the output unit to output the area of the target object (e.g., a passenger car 201 or a truck 202) corresponding to the input character string or the selected thread when a character string (e.g., one of "a passenger car" and "a truck") is input via the input interface 140, or when one of the threads is selected via the input interface 140. That is, the output control unit 114 may switch between displaying and hiding multiple threads via the input interface 140.

The output control unit 114 may perform control to output a region of the target object having a size equal to or larger than an input value (threshold value). The output control unit 114 may perform control to identify a region of the target object having a value (region size) equal to or larger than a threshold value based on the value (region size) assigned by the estimation unit 113 and the threshold value input via the input interface 140, and to output the identified region of the target object.
As a specific example, the output control unit 114 may compare the size of the polygons of each object (each target object) with a threshold value input via the input interface 140, and perform filtering to hide each object (each target object) corresponding to a polygon whose size is less than the threshold value.

As an example, as shown in FIG. 4, the output control unit 114 may display a slider 143 on the input interface 140 for specifying the threshold in the range of 0 to 100, and when the slider 143 is moved in response to the operation of the input unit 121, the threshold may be changed in response to the position of the slider 143. Based on the threshold and the size of the area of the object (target object) (a passenger car 201 in the example shown in FIG. 4), the output control unit 114 may display the area of the object (target object) (a passenger car 201 in the example shown in FIG. 4) that is equal to or greater than the threshold on the display unit 133 in a manner different from other areas. In the example shown in FIG. 4, the output control unit 114 shades the area of the target object (a regular car 201) in the image 200, and does not shade other areas (e.g., a truck 202 and a regular car of a size less than the threshold, not shown in FIG. 4), so that the area of the target object (a regular car 201) is shown in a different manner from the other areas (e.g., the truck 202, etc.).

The output control unit 114 may perform control so as to add an annotation based on the instruction content 141 to the region of the target object and output the result. The output control unit 114 may also perform control so as to add an annotation to the region of the target object identified in the image 200 based on the character string (instruction content 141 (prompt)) received by the reception unit 112 as described above. The output control unit 114 may perform control so as to add an annotation based on the instruction content 141 (prompt) to the region of the target object having a size equal to or larger than the value (threshold value) input via the input interface 140 and output the result.
As an example, in the case shown in Figure 4, the output control unit 114 may identify the area of the target object (standard car 201) by adding a diagonal line based on the character string (instruction content 141 (prompt)) "standard car", and may annotate the area of the target object (standard car 201) with "standard car".

As an example of the above-mentioned output, the output control unit 114 may control the communication unit 131 to transmit at least one piece of information of the region of the target object and the group of the annotated images 200 to an external device (not shown). The external device may be, for example, a server, a user terminal, or the like.
As another example of the output, the output control unit 114 may control the storage unit 132 to store at least one piece of information on the region of the target object and the group of annotated images 200 .
As an example of output, the output control unit 114 may control the display unit 133 to display at least one of the region of the target object and the group of the annotated image 200. In this case, the output control unit 114 may control the display unit 133 to display the region of the target object in the image 200 in a manner different from other regions excluding the region (e.g., a different color, a different line thickness surrounding the outline, or blinking or brightening the region of the target object).

The learning unit 115 may generate a trained model by, for example, learning an image or the like that has been annotated by the output control unit 114 with a character string (the instruction content 141 (prompt)). That is, when an annotation is added to a region of a target object, the learning unit 115 may learn the region of the target object to generate a trained model.
When using the example described above, i.e., when there is an area annotated with a passenger car 201 and an area annotated with a truck 202 as objects recorded in an image, the learning unit 115 may learn the images with those annotations to generate a learned model.

The AI unit 116 may input a target to the trained model generated by the learning unit 115 and estimate a target object in the target. An example of the target may be an image (image information) such as a still image or a video. That is, when the AI unit 116 inputs an image (target) to be estimated to the trained model, it becomes possible to estimate a target object (estimated object) recorded in the image (target).
As a specific example, when the learning unit 115 generates a trained model that has learned the area of passenger car 201 annotated with passenger car 201 and the area of truck 202 annotated with truck 202, the AI unit 116 may estimate passenger car 201 or truck 202 (target object) in the image (target) based on the image (target) and the trained model.

[Information processing method]
Next, an information processing method according to an embodiment will be described.
FIG. 5 is a flowchart illustrating an information processing method according to an embodiment.

In step ST101, the acquisition unit 111 acquires an image 200 (image information).

In step ST102, the reception unit 112 receives an instruction content 141 (prompt) for the target object. The reception unit 112 may receive an instruction content 141 (prompt) for the target object in the form of a character string (text).

In step ST103, the estimation unit 113 estimates the area of the target object recorded in the image 200 acquired in step ST102 based on the instruction content 141 received in step ST101. The estimation unit 113 may estimate at least one of the pixel area of the target object in the image 200 and the contour boundary of the target object as the area of the target object. The estimation unit 113 may assign a value according to the size of the area of the target object.

In step ST104, the output control section 114 controls the output section to output the area of the target object estimated in step ST103.
The output control unit 114 may control the output of an area of the target object that is equal to or larger than the input threshold (value) based on the value assigned in step ST103 (a value corresponding to the size of the area of the target object) and a threshold (value) input via the input interface 140.
The output control section 114 may annotate the region of the target object based on the instruction content 141 (prompt such as a character string) received in step ST102.

When an annotation is added to the region of the target object in accordance with the processing of step ST104, the learning unit 115 may learn the region of the target object to generate a learned model.
The AI unit 116 may input an object to the trained model generated by the learning unit 115 and estimate a target object within the object.

[Functions and circuits]
Next, the functions and circuits of the information processing device 100 will be described.
Each unit of the information processing device 100 may be realized as a function of a computer's arithmetic processing device, etc. That is, the acquisition unit 111, the reception unit 112, the estimation unit 113, the output control unit 114, the learning unit 115, and the AI unit 116 (control unit 110) of the information processing device 100 may be realized as an acquisition function, a reception function, an estimation function, an output control function, a learning function, and an AI function (control function) by the computer's arithmetic processing device, etc.
The information processing program can cause a computer to realize each of the above-mentioned functions. The information processing program may be recorded in a non-transitory computer-readable storage medium, such as a memory, a solid state drive, a hard disk drive, or an optical disk. The storage medium may be rephrased as a non-transitory computer-readable medium that stores the information processing program. The information processing program may also be transmitted online.
Also, as described above, each part of the information processing device 100 may be realized by a computer arithmetic processing device or the like. The arithmetic processing device or the like is configured by, for example, an integrated circuit or the like. Therefore, each part of the information processing device 100 may be realized as a circuit constituting the arithmetic processing device or the like. That is, the acquisition unit 111, the reception unit 112, the estimation unit 113, the output control unit 114, the learning unit 115, and the AI unit 116 (control unit 110) of the information processing device 100 may be realized as an acquisition circuit, a reception circuit, an estimation circuit, an output control circuit, a learning circuit, and an AI circuit (control circuit) constituting the arithmetic processing device or the like of a computer.
The input unit 121, the communication unit 131, the storage unit 132, and the display unit 133 (output unit) of the information processing device 100 may be realized as, for example, an input function including the functions of an arithmetic processing device, and a communication function, a storage function, and a display function (output function). The input unit 121, the communication unit 131, the storage unit 132, and the display unit 133 (output unit) of the information processing device 100 may be realized as, for example, an input circuit, a communication circuit, a storage circuit, and a display circuit (output circuit) by being configured by an integrated circuit, etc. The input unit 121, the communication unit 131, the storage unit 132, and the display unit 133 (output unit) of the information processing device 100 may be realized as, for example, an input unit device, a communication device, a storage device, and a display device (output device) by being configured by a plurality of devices.

The information processing device 100 can combine one or any two or more of the above-mentioned units.
In this disclosure, the term "information" is used, but the term "information" can be replaced with "data" and the term "data" can be replaced with "information."

[Aspects and Effects of the Present Embodiment]
Next, one aspect of this embodiment and the effects of each aspect will be described. Note that each aspect described below is an example at the time of filing, and this embodiment is not limited to the aspects described below. In other words, this embodiment is not limited to the aspects described below, and may be realized by appropriately combining each of the above-mentioned parts. In addition, a lower aspect may be able to cite any of the higher aspects.
In addition, the effects of the present embodiment described below are merely examples, and the effects of each aspect are not limited to those described below. In addition, each aspect may, for example, achieve at least one of the effects described below.

(Aspect 1)
An information processing device of one embodiment includes an acquisition unit that acquires an image, a reception unit that receives instructions regarding a target object, an estimation unit that estimates the area of the target object to be recorded in the image acquired by the acquisition unit based on the instructions received by the reception unit, and an output control unit that controls the area of the target object estimated by the estimation unit to be output.
As a result, when the information processing device receives an instruction (prompt) in the form of a string of characters (text), it can output (e.g., display) the area of the target object in the image according to the instruction (prompt).
For example, when there are areas of multiple types of target objects in an image, the information processing device can output (e.g., display) the area of one type of target object by switching the output content (display content).
The information processing device can add an annotation corresponding to a character string (text) to a region of a target object to be output.

(Aspect 2)
In the information processing device of one aspect, the estimation unit may estimate, as the region of the target object, at least one of a pixel region of the target object in the image and a contour boundary of the target object.
This allows the information processing device to annotate a target object in an image based on at least one of a group of pixel regions and a group of contour boundaries.

(Aspect 3)
In one embodiment of the information processing device, the estimation unit may assign a value according to the size of the area of the target object, and the output control unit may control the output to output an area of the target object that is equal to or larger than the input value.
This allows the information processing device to estimate that a region whose size is less than the input value (threshold value) is noise, and prevents the information processing device from annotating a wrong region (noise) of a target object.

(Aspect 4)
In the information processing device of one aspect, the receiving unit may receive instructions in the form of text regarding the target object.
This allows the information processing device to annotate a region of the target object that corresponds to the content of the text (character string) based on the instruction content (prompt) of the text (character string). That is, the information processing device can annotate a region of the target object that corresponds to the content of the text (character string).

(Aspect 5)
In the information processing device of one aspect, the output control unit may perform control so as to add an annotation based on the instruction content to the region of the target object and output the region.
This allows the information processing device to generate a trained model by learning, for example, an annotated region of a target object in an image. That is, the information processing device can automatically annotate a target object in an image and generate ground truth data.
That is, when the information processing device receives an upload of an image and further receives instructions (prompts) regarding the target object via the input unit, it can automatically annotate the target object in the image. Since the information processing device can automatically annotate, it is possible to significantly reduce the work time and work man-hours (work costs) compared to manually identifying the target object in the image and annotating it.

(Aspect 6)
An information processing device according to one embodiment may include a learning unit that, when an annotation is added to a region of a target object, learns the region of the target object and generates a learned model.
This enables the information processing device to automatically perform learning based on the correct answer data and generate a learned model.

(Aspect 7)
An information processing device according to one embodiment may include an AI unit that inputs an object to a trained model generated by a learning unit and estimates a target object among the objects.
This enables the information processing device to make inferences about various target objects that were learned when generating the trained model.

(Aspect 8)
In one aspect of the information processing method, a computer executes an acquisition step of acquiring an image, a reception step of receiving instructions regarding a target object, an estimation step of estimating the area of the target object recorded in the image acquired by the acquisition step based on the instructions received by the reception step, and an output control step of controlling the output of the area of the target object estimated by the estimation step.
As a result, the information processing method can achieve the same effects as the information processing device according to the above-described aspect.

(Aspect 9)
One embodiment of an information processing program enables a computer to realize an acquisition function for acquiring an image, a reception function for receiving instructions regarding a target object, an estimation function for estimating the area of the target object recorded in the image acquired by the acquisition function based on the instructions received by the reception function, and an output control function for controlling the output of the area of the target object estimated by the estimation function.
As a result, the information processing program can achieve the same effects as the information processing device according to the above-described aspect.

100 Information processing device 110 Control unit 111 Acquisition unit 112 Reception unit 113 Estimation unit 114 Output control unit 115 Learning unit 116 AI unit 121 Input unit 131 Communication unit 132 Storage unit 133 Display unit 140 Input interface 141 Instruction content (prompt)
142 Chat reply content 143 Slider 144 Chat reply content 200 Image 201 Regular car 202 Truck

Claims (9)

An acquisition unit that acquires an image;
a reception unit that receives instructions about a target object by inputting and replying to the inputted character string through chat ;
an estimation unit that estimates an area of a target object recorded in an image acquired by the acquisition unit based on the instruction content accepted by the acceptance unit;
an output control unit that controls the area of the target object estimated by the estimation unit to be output;
An information processing device comprising: The information processing apparatus according to claim 1 , wherein the estimation unit estimates, as the region of the target object, at least one of a pixel region of the target object in the image and a contour boundary of the target object. The estimation unit assigns a value according to a size of a region of the target object,
The information processing apparatus according to claim 1 , wherein the output control unit performs control so as to output a region of the target object having a size equal to or larger than an input value. The information processing apparatus according to claim 1 , wherein the output control unit performs control so as to add an annotation based on an instruction content to a region of the target object and output the region. A learning unit that learns the region of the target object and generates a learned model when the region of the target object is annotated.
The information processing device according to any one of claims 1 to 4 . The learning unit includes an AI unit that inputs an object to the trained model generated by the learning unit and estimates a target object in the object.
The information processing device according to claim 5 . An acquisition unit that acquires an image;
A reception unit that receives instructions regarding a target object;
an estimation unit that estimates an area of a target object recorded in the image acquired by the acquisition unit based on the instruction content accepted by the acceptance unit, and assigns a value according to the size of the area of the target object;
an output control unit that controls to output a region of the target object estimated by the estimation unit, the region being equal to or larger than an input value;
An information processing device comprising: The computer
An acquisition step of acquiring an image;
a receiving step of receiving an instruction content for a target object by an input character string by performing input and a reply using a chat ;
an estimation step of estimating an area of a target object recorded in the image acquired by the acquisition step based on the instruction content accepted by the acceptance step;
an output control step of controlling to output the area of the target object estimated by the estimation step;
An information processing method for performing the above. On the computer,
an acquisition function for acquiring an image;
A reception function for receiving instructions regarding a target object by inputting and replying to the inputted character string through chat ;
an estimation function that estimates an area of a target object recorded in an image acquired by the acquisition function based on the instruction content accepted by the acceptance function;
an output control function that controls to output the area of the target object estimated by the estimation function;
An information processing program that realizes this.

Images