Information processing device, information processing method, and program

The information processing device uses reinforcement learning to identify objects in images without positional annotations, enhancing determination accuracy by calculating confidence levels and providing rewards, addressing the inefficiencies of existing image recognition methods.

JP7593605B1Active Publication Date: 2024-12-03FAST ACCOUNTING INC
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
JP2023572239
Authority / Receiving Office
JP · JP
Patent Type
Patents
Current Assignee / Owner
Filing Date
2023-11-21
Publication Date
2024-12-03
Estimated Expiration
2043-11-21

AI Technical Summary

Technical Problem

Existing image recognition techniques require time-consuming and effort-intensive annotations to indicate the position of target objects in images.

Method used

An information processing device and method that utilizes reinforcement learning to identify objects in images without relying on positional annotations by using an agent to select operations on regions of interest, calculating certainty levels, and providing rewards based on confidence changes to determine the object's type and position.

Benefits of technology

Enables accurate identification of objects in images without prior positional information, particularly suitable for logo recognition tasks where objects vary in size and position, improving determination accuracy.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 0007593605000003
    Figure 0007593605000003
  • Figure 0007593605000004
    Figure 0007593605000004
  • Figure 0007593605000005
    Figure 0007593605000005
Patent Text Reader

Abstract

The information processing device 1 has an acquisition unit 131 that acquires image data including an object to be judged as a subject, a selection unit 132 that allows an agent operating an area including at least a portion of the image data for identifying the object to be judged to select an operation to change the area, a calculation unit 133 that calculates a certainty factor for the type of object to be judged included in the area after the operation selected by the agent is performed, a reward determination unit 134 that gives the agent a reward determined based on the certainty factor calculated by the calculation unit, and an output unit 135 that outputs, when the agent selects an operation to determine the area in which the object to be judged is located, the position of the area at the time the determination operation was performed and the type of object to be judged in the image data.
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

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

[0002] There is known a technique for identifying an object included as a subject in an image (for example, Non-Patent Document 1). [Prior art documents] [Non-patent literature]

[0003] [Non-Patent Document 1] Juan C Caicedo and Svetlana Lazebnik, “Active object localization with deep reinforcement learning,” in Proceedings of the IEEE International Conference on Computer Vision, 2015, pp. 2488-2496. Summary of the Invention [Problem to be solved by the invention]

[0004] Existing image recognition technologies require annotations that indicate the position of the object to be determined contained in the image, but there is a problem in that preparing annotations that indicate the position in the image is time-consuming.

[0005] Therefore, the present invention has been made in consideration of these points, and aims to make it possible to identify an object to be determined contained in an image without relying on an annotation indicating the position of the object to be determined. [Means for solving the problem]

[0006] An information processing device of a first aspect of the present invention has an acquisition unit that acquires image data including an object to be judged as a subject, a selection unit that allows an agent manipulating an area including at least a portion of the image data to select an operation to change the area to identify the object to be judged, a calculation unit that calculates a certainty of the type of the object to be judged included in the area after the operation selected by the agent is performed, a reward determination unit that gives the agent a reward determined based on the certainty calculated by the calculation unit, and an output unit that, when the agent selects an operation to determine the area in which the object to be judged is located, outputs the position of the area at the time of the determination operation and the type of the object to be judged in the image data.

[0007] The calculation unit may calculate a probability that the type of the object to be judged belongs to each of a plurality of predetermined classes based on the operation selected by the agent, and calculate a certainty of the type of the object to be judged included in the area based on the calculated probability.

[0008] The selection section may cause the agent to select an operation to be performed on the area based on an operation previously selected by the agent and an image feature amount extracted from the image data.

[0009] The reward determination unit may give a positive reward to the agent when the confidence level of the agent when performing the selected operation is higher than the confidence level before performing the operation, and may give a negative reward to the agent when the confidence level of the agent when performing the determined operation is lower than the confidence level before performing the operation.

[0010] The reward determination unit may give a positive reward to the agent when the agent determines a final position of the object to be judged in the image data and the confidence level in the final position of the object to be judged is greater than a threshold value.

[0011] The image processing device may further include a learning unit that uses training image data, which is image data for learning that includes the object to be determined as a subject, and the type of the object to be determined contained in the training image data as teacher data to teach the agent an operation of determining the area in which the object to be determined is located.

[0012] An information processing method of a second aspect of the present invention includes the steps of: acquiring image data including an object to be judged as a subject, executed by a computer; having an agent that operates an area including at least a portion of the image data to identify the object to be judged select an operation to change the area; calculating a certainty of the type of the object to be judged included in the area after the operation selected by the agent is performed; giving the agent a reward determined based on the calculated certainty; and, when the agent selects an operation to determine the area in which the object to be judged is located, outputting the position of the area at the time of the determination operation and the type of the object to be judged in the image data.

[0013] In a third aspect of the program of the present invention, a computer is caused to execute the steps of acquiring image data including an object to be judged as a subject, having an agent that operates an area including at least a portion of the image data to identify the object to be judged select an operation to change the area, calculating a degree of certainty of the type of the object to be judged included in the area after the operation selected by the agent is performed, giving the agent a reward determined based on the calculated degree of certainty, and when the agent selects an operation to determine the area in which the object to be judged is located, outputting the position of the area at the time of the determination operation and the type of the object to be judged in the image data. Effect of the Invention

[0014] According to the present invention, it is possible to determine the type of an object contained in an image even when there is no information on the position of the object in the image. [Brief description of the drawings]

[0015] [Figure 1] 1 is a diagram for explaining an overview of an information processing system S. [Diagram 2] 1 is a block diagram showing a configuration of an information processing device 1. FIG. [Diagram 3] 2 is a diagram illustrating a schematic diagram of processing in the information processing device 1. FIG. [Figure 4] 13 is a diagram showing an example of a screen displayed by an output unit 135. FIG. [Diagram 5] 3 is a flowchart showing a process flow in the information processing device 1. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0016] [Outline of Information Processing System S] 1 is a diagram for explaining an overview of an information processing system S. The information processing system S is a system for identifying an object included in an image. The information processing system S includes an information processing device 1 and an information terminal 2.

[0017] The information processing device 1 is a device that determines the type of object included in an image and the position where the object is included. More specifically, the information processing device 1 acquires an image, and outputs the position of the object included in the acquired image and the class of the object. In this specification, the class indicates the type of the object to be determined. For example, if the object to be determined is a logo for identifying a company or service, the class indicates the type of logo. The logo may indicate a business that provides a product or the like, or may represent a product or the like provided by the business. The information processing device 1 is, for example, a server. The number of classes is determined based on the type of the object to be determined included in the image data provided as teacher data.

[0018] The information terminal 2 is a terminal operated by a user. The information terminal 2 is, for example, a smartphone, a tablet, or a personal computer. The information terminal 2 transmits an image to be judged to the information processing device 1, acquires a result of the judgment made by the information processing device 1, and displays the acquired judgment result on a display unit.

[0019] The processing in the information processing system S will be described with reference to Fig. 1(a). The information processing device 1 acquires image data D ((1) in Fig. 1). The image data D includes an object to be determined as a subject. The information processing device 1 identifies the position and type of the object to be determined included in the acquired image data D ((2) in Fig. 1). The information processing device 1 outputs the identified position and type of the object to be determined to the information terminal 2 ((3) in Fig. 1).

[0020] The process for identifying the position and type of the object to be determined will be described with reference to FIG. 1(b). The information processing device 1 uses a reinforcement learning technique as shown below to have the agent A select an operation to be performed on the region R, and identifies the position and type of the object in the image. The region R is a region for identifying the object to be determined in the input image, and is a so-called bounding box. The agent A determines an operation to be performed on the region R based on the feature amount of the image data D. The agent A is a trained model that has learned the operation to be performed on the region R so as to maximize the reward given according to the operation. As an example, the agent A is a trained model generated by training using a known DQN (Deep Q Network).

[0021] Agent A selects an operation to be performed on region R based on the features of the image contained in region R in image data D ((2-1) in Figure 1). Agent A predicts the reward to be given when performing each operation based on the features of the image contained in region R in image data D, and selects an operation based on the predicted reward. As an example, agent A selects an operation that maximizes the reward. The operation to be performed on region R is either enlarging, reducing, changing the aspect ratio of region R, moving vertically or horizontally, or ending the operation. In other words, when agent A performs an operation on region R, the position, size, and shape of region R are changed. Note that Figure 2(b) shows an example in which an operation to reduce region R is performed.

[0022] The information processing device 1 calculates a confidence score for each class of the classification target based on the feature amount of the image included in the region R after the operation ((2-2) in FIG. 1). When an image is input, the information processing device 1 calculates a prediction probability using a class prediction model that has been trained to output a prediction probability for each class, and calculates a confidence score based on the calculated prediction probability. The prediction probability indicates the probability that the object to be judged is predicted to belong to the class. The confidence score is an index that indicates the uncertainty of the prediction. When the prediction probabilities of multiple classes are high, the confidence score is small, and when only the prediction probability of a specific class is high, the confidence score is large. The information processing device 1 determines a reward to be given to agent A based on the calculated confidence score. A method for determining the reward will be described later.

[0023] When the information processing device 1 acquires an image of an object to be determined, it repeats the above-mentioned processes (2-1) to (2-3) until agent A selects to end the operation, thereby identifying the position and type of the object to be determined.

[0024] If region R does not properly capture the object to be determined contained in the image, the certainty factor is considered to be small. In other words, if the certainty factor is large, region R is considered to specify an appropriate position. Therefore, in the information processing device 1, by using the certainty factor, it is possible to determine the type of the object to be determined contained in the image even when there is no annotation indicating the position of the object to be determined. In particular, the information processing device 1 is suitable for logo recognition tasks in which logos contained in images vary in size and position.

[0025] [Configuration of information processing device 1] 2 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 selection unit 132, a calculation unit 133, a reward determination unit 134, an output unit 135, and a learning unit 136.

[0026] The information processing device 1 will be described in detail with reference to Fig. 3. Fig. 3 is a diagram showing a schematic diagram of processing in the information processing device 1. The acquisition unit 131 acquires image data D including an object to be determined as a subject. The acquisition unit 131 acquires the image data D from the information terminal 2. The object to be determined is an object to be classified and located by image recognition. As an example, the object to be determined is a logo mark for identifying a company or a service provided by the company. In other words, the image data D is image data including an object to which the logo mark is attached as a subject.

[0027] The acquisition unit 131 inputs the acquired image data D to an image encoder and extracts image features. When the image data D is acquired, the acquisition unit 131 may set the region R at a predetermined position (for example, a region including the entire image) that is set in advance as an initial position. The acquisition unit 131 may cause the calculation unit 133 to calculate the confidence level at the time when the image data D is acquired, as described below.

[0028] The selection unit 132 causes the agent A, which operates an area including at least a part of the image data D and for identifying an object to be determined, to select an operation to change the area R. The selection unit 132 inputs a feature amount included in the image to the agent A, and causes the agent A to output information indicating the operation content to be performed on the area R.

[0029] The calculation unit 133 calculates the certainty of the type of the object to be determined included in the region R in which the operation selected by the agent A was performed. The calculation unit 133 calculates the probability that the object included in the region R of the image data D belongs to each of a plurality of classes based on the operation selected by the agent A, and calculates the certainty of the type of the object to be determined included in the region R based on the calculated probability. As an example, the storage unit 12 stores a class prediction model that is a trained model trained to output a predicted probability for each class when a feature amount of an image is input, and the calculation unit 133 inputs the feature amount of the image into the class prediction model to output a predicted probability for each class and calculates the certainty. The calculation unit 133 calculates the certainty for each class that can be classified based on the feature amount of the image included in the region R. The calculation unit 133 may normalize the certainty of each class.

[0030] The reward determination unit 134 gives the reward determined based on the confidence calculated by the calculation unit 133 to the agent A. The reward determination unit 134 determines the reward to be given to the agent A based on the confidence based on the area R before the operation and the confidence based on the area R after the operation. As an example, the reward determination unit 134 may give a positive reward to the agent A when the confidence based on the area after the operation is higher than the confidence based on the area before the operation, and may give a negative reward to the agent A when the confidence based on the area after the operation is lower than the confidence based on the area before the operation.

[0031] The reward determination unit 134 determines the reward based on the following formula, for example. Here, Re represents the reward, C' represents the confidence of the target class based on the region R after the operation, and C represents the confidence of the target class based on the region R before the operation. Note that sign(C'-C) is a sign function that is -1 when the sign of the argument is negative, +1 when the sign of the argument is positive, and 0 when the argument is 0. The target class is, for example, the class with the highest confidence among the classes to be determined.

[0032]

number

[0033] The selection unit 132, the calculation unit 133, and the reward determination unit 134 repeat the above-mentioned process until the end of the operation is selected by the agent A. When the agent A selects an operation to determine the area in which the object to be determined is located, the output unit 135 outputs the position of the area R at the time of the determination operation and the type of the object to be determined in the image data D. The output unit 135 may display on the information terminal 2 a screen including the position of the area R at the time the end of the operation is selected and the class of the determination result. FIG. 4 is a diagram showing an example of a screen displayed by the output unit 135. In the screen shown in FIG. 4, an object indicating the position of the area R at the time the end of the operation is selected is superimposed on the image P of the object to be determined, and information indicating the class of the determination result is displayed. The class of the determination result is, for example, a class with the highest degree of certainty at the time the end of the operation is selected.

[0034] The learning unit 136 updates the parameters of the agent A based on the reward predicted by the agent A and the reward actually given to the agent A as a result of repeating the operation, and causes the agent A to learn.

[0035] In addition, in the processing at the inference stage when the learning of agent A is completed, the processing of the reward determination unit 134 may be omitted.

[0036] By configuring the information processing device 1 in this way, it is possible to determine the type of an object to be determined contained in an image even when there is no information on the position of the object to be determined in the image.

[0037] The selection unit 132 may be configured to determine an operation by further using a history of past operations as input data. That is, the selection unit 132 causes the agent A to select an operation to be performed on the region R based on the operation selected immediately before by the agent A and the feature amount of the image extracted from the image data D. The selection unit 132 may also cause the agent A to select an operation to be performed on the region R based on the contents of the operations performed several times in the past. By configuring the selection unit 132 in this way, it is possible to select a more appropriate operation content, thereby improving the accuracy of the determination.

[0038] The reward determination unit 134 may be configured to determine a reward to be given to agent A based on the confidence level when the operation on region R is completed and the final position of region R is determined. When agent A determines the final position of the object to be determined in image data D, the reward determination unit 134 gives a positive reward to agent A if the confidence level for the final position of the object to be determined is greater than a threshold value. As an example, the reward when the final position of region R is determined is expressed by the following formula.

[0039]

number

[0040] By configuring the reward determination unit 134 in this way, it is possible to achieve a balance between the reward based on the determined operation and the reward based on the final position, thereby improving the accuracy of the determination.

[0041] [About learning] The information processing device 1 may be configured to perform agent learning in two stages. First, the acquisition unit 131 acquires learning image data and a class to which a subject included in the learning image data belongs. The learning image data is image data for learning that includes a determination target as a subject. The learning unit 136 uses the learning image data and the class of the determination target included in the learning image data as teacher data to have the agent A learn an operation to determine an area in which the determination target is located.

[0042] The learning unit 136 trains a class prediction model using the training image data and the class of the object to be determined contained in the training image data. As an example, the learning unit 136 may update parameters of the class prediction model based on the cross entropy error between the class predicted by the class prediction model and the correct class. At this stage, the learning unit 136 trains the class prediction model using training image data in which the object to be determined is largely reflected in the image.

[0043] Next, the learning unit 136 simultaneously trains the agent A and the class prediction model based on the training image data and the class of the object to be determined contained in the training image data. As an example, the learning unit 136 updates the parameters of the agent A based on the reward predicted by the agent A and the reward actually given to the agent A.

[0044] By configuring the learning unit 136 to perform learning in two stages in this way, it is possible to efficiently and comprehensively learn information about the object to be determined and the optimal operation.

[0045] [Processing flow in information processing device 1] Fig. 5 is a flowchart showing the flow of processing in the information processing device 1. The flowchart shown in Fig. 5 starts from the point in time when the information processing device 1 receives an instruction to execute a determination process.

[0046] The acquisition unit 131 acquires image data D (S01). The acquisition unit 131 inputs the image data D to an image encoder and extracts image features (S02). The acquisition unit 131 determines the initial position of a region R (S03).

[0047] The calculation unit 133 calculates the confidence level of each class based on the image feature amount and the position of the region R (S04). Specifically, the calculation unit 133 calculates the predicted probability that the object to be determined included in the image data D belongs to each class based on the extracted feature amount, and calculates the confidence level based on the calculated predicted probability.

[0048] The selection unit 132 inputs the feature amount of the image and the position of the region R to the agent A, and causes the agent A to select an operation for the region R (S05). The selection unit 132 may further input information indicating past operations to the agent A, and cause the agent A to output an operation.

[0049] The selection unit 132 determines whether or not an end condition is satisfied (S06). The end condition is that the agent A has selected to end the operation. If the end condition is satisfied (YES in S06), the information processing device 1 advances the process to S10.

[0050] If the termination condition is not satisfied (NO in S06), the calculation unit 133 calculates the confidence level of each class based on the feature amount of the image and the position of the region R after the operation (S07). The reward determination unit 134 determines the reward to be given to the agent A based on the confidence level based on the region R before the operation and the confidence level based on the region R after the operation (S08).

[0051] The information processing device 1 determines whether or not the processes of S05 to S08 have been executed a preset number of times (S09). If the number of times that the processes of S05 to S08 have been executed is equal to or greater than the preset number of times (YES in S09), the information processing device 1 advances the process to S10. If the preset number of times has not been reached (NO in S09), the information processing device 1 advances the process to S05. Note that instead of or in addition to whether or not the processes of S05 to S08 have been executed a preset number of times, the information processing device 1 may advance the process to S10 when the reward has reached a certain value.

[0052] If the termination condition is satisfied (YES in S06), or if the number of times the processes of S05 to S08 have been executed is equal to or greater than a predetermined number of times (YES in S09), the calculation unit 133 determines a reward to be given to agent A based on the confidence level when the final position of region R is determined (S10). The output unit 135 outputs the determination result (S11). Specifically, the output unit 135 outputs the type of the object to be determined and the position of the object to be determined based on the prediction result calculated by the calculation unit 133 and the final position of region R determined by agent A. Then, the information processing device 1 ends the process.

[0053] [Effects of this embodiment] As described above, the information processing device 1 can determine the type of an object to be determined contained in an image even when there is no information on the position of the object to be determined in the image.

[0054] Although the present invention has been described above using the embodiments, 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 distributing or integrating functionally or physically in any unit. In addition, new embodiments resulting from any combination of multiple embodiments are also included in the embodiments of the present invention. The effect of the new embodiment resulting from the combination combines the effect of the original embodiment. [Explanation of symbols]

[0055] 1. Information processing device 2. Information terminal 11 Communications Department 12 Storage section 13 Control section 131 Acquisition Department 132 Selection section 133 Calculation section 134 Remuneration Determination Department 135 Output section 136 Learning Department

Claims

1. An acquisition unit that acquires image data including an object to be determined as a subject; a selection unit that causes the agent to select an operation for changing the area, the agent operating the area including at least a portion of the image data to identify the object to be determined, the selection unit predicting a reward to be given when the agent operates the area and selecting an operation for the area that maximizes the predicted reward; and a calculation unit that calculates a certainty factor of the type of the object to be determined that is included in the area after the agent has performed an operation selected by the agent; a reward determination unit that gives a reward to the agent, the reward being determined based on the confidence calculated by the calculation unit; a learning unit that causes the agent to learn an operation to be performed on the area based on the reward determined by the reward determination unit and the reward predicted by the agent; an output unit that, when the agent selects an operation to determine an area in which the object to be determined is located, outputs a position of the area at the time of the operation to determine and a type of the object to be determined in the image data; An information processing device having the above configuration.

2. the calculation unit calculates a probability that the type of the object to be determined belongs to each of a plurality of predetermined classes based on the operation selected by the agent, and calculates a certainty of the type of the object to be determined included in the area based on the calculated probabilities. The information processing device according to claim 1 .

3. the selection unit causes the agent to select an operation to be performed on the region based on an operation previously selected by the agent and an image feature amount extracted from the image data; The information processing device according to claim 1 .

4. The remuneration determination unit, giving a positive reward to the agent when the confidence level of the agent performing the selected operation is higher than the confidence level of the agent prior to performing the operation; giving a negative reward to the agent when the confidence level after performing the determined operation is lower than the confidence level before performing the operation; The information processing device according to claim 3 .

5. the reward determination unit, when the agent determines a final position of the object to be determined in the image data, gives a positive reward to the agent if the degree of confidence in the final position of the object to be determined is greater than a threshold value; 5. The information processing device according to claim 3.

6. A learning unit is further provided for making the agent learn an operation for determining an area in which the object to be determined is located, using training image data, which is image data for learning including an object to be determined as a subject, and a type of the object to be determined included in the training image data as teacher data. The information processing device according to claim 1 .

7. The computer executes A step of acquiring image data including an object to be determined as a subject; a step of having an agent that manipulates an area including at least a portion of the image data for identifying the object to be determined predict a reward that will be given when the agent manipulates the area and selects an operation for the area that maximizes the predicted reward by selecting an operation for the area; and calculating a certainty factor of the type of the object to be determined that is included in the area after the agent performs the selected operation; A step of giving a reward to the agent, the reward being determined based on the calculated confidence level; having the agent learn an operation to be performed on the area based on the determined reward and a reward predicted by the agent; a step of outputting, when the agent selects an operation for determining an area in which the object to be determined is located, a position of the area at the time of performing the operation for determining and a type of the object to be determined in the image data; An information processing method comprising the steps of:

8. On the computer, A step of acquiring image data including an object to be determined as a subject; a step of having an agent that manipulates an area including at least a portion of the image data for identifying the object to be determined predict a reward that will be given when the agent manipulates the area and selects an operation for the area that maximizes the predicted reward by selecting an operation for the area; and calculating a certainty factor of the type of the object to be determined that is included in the area after the agent performs the selected operation; A step of giving a reward to the agent, the reward being determined based on the calculated confidence level; having the agent learn an operation to be performed on the area based on the determined reward and a reward predicted by the agent; a step of outputting, when the agent selects an operation for determining an area in which the object to be determined is located, a position of the area at the time of performing the operation for determining and a type of the object to be determined in the image data; A program that executes the following.

Citation Information

Patent Citations

  • Method, system, and computer program

    WO2022124224A1

  • Image processing device, image processing system, image processing method, and program

    WO2022255418A1