CN-114788265-B - Information processing method, information processing system, and storage medium

CN114788265BCN 114788265 BCN114788265 BCN 114788265BCN-114788265-B

Abstract

As an information processing method executed by a computer, a plurality of inference results are acquired as a result of inferences made on the same input data by a plurality of inference engines, respectively (S11), the influence of the input data on the acquired plurality of inference results is acquired for each inference result (S12), one or more combinations are determined based on the plurality of inference results, the one or more combinations being combinations of the plurality of inference results with each other (S14), and the influence acquired for each of the inference results included in the same combination among the determined plurality of inference results is presented in parallel or overlapped with each other via a presentation means (S15).

Inventors

FUJIMURA RYOTA
Ishii ikunori
Open up also at mountain pass
URABE HIROAKI

Assignees

松下知识产权经营株式会社

Dates

Publication Date: 20260508
Application Date: 20200918
Priority Date: 20191202

Claims (11)

1. An information processing method, executed by a computer, wherein, For each of a plurality of inference engines, obtaining a plurality of object detection results for the same image by the inference engine, Obtaining, for each of the plurality of object detection results of the respective inference engines, influence information indicating the presence or absence of an influence of the image on each of the plurality of object detection results and, if the influence is present, information indicating one or both of the magnitude and the directivity of the influence, Determining one or more combinations of object detection results of the same category among the plurality of object detection results of the respective inference engines based on the categories of the detected objects of the plurality of object detection results of the respective inference engines, The influence information acquired for each of the object detection results included in the same combination among the plurality of object detection results of the inference engine is superimposed on each of the object detection results included in the same combination via a presentation device, and then the object detection results included in the same combination, on which the influence information is superimposed, are presented in parallel with each other.
2. The information processing method according to claim 1, wherein, Each of the plurality of inference engines is an object detector.
3. The information processing method according to claim 1, wherein, The object detection results respectively comprise a detection frame, The one or more combinations are also determined based on the coincidence or positional relationship of the detection frames included in the plurality of object detection results of the respective inference engines.
4. The information processing method according to claim 3, wherein, The plurality of object detection results each include a likelihood of detection, The one or more combinations are also determined based on the proximity of the likelihoods of the detections included in the plurality of object detection results of the respective inference engines to each other.
5. The information processing method according to claim 2, wherein, In the presenting of the influence information, when the plurality of object detection results generated by any one of the plurality of object detectors do not include an object detection result belonging to a category of any one of the one or more combinations, the substitute data for the object detection result is presented in parallel.
6. The information processing method according to claim 2, wherein, Further, when there is an isolated object detection result that is not included in any one of the one or more combinations among the plurality of object detection results, the influence information and the substitution data acquired for the isolated object detection result are presented in parallel.
7. The information processing method according to claim 1, wherein, An operation of selecting the image is also accepted, The image is switched to the image selected by the operation.
8. The information processing method according to claim 2, wherein, An operation of selecting a combination of the object detection results is also accepted, The object detection result or the presented object detection result opposed to the presented influence information is switched to the combined object detection result selected by the operation.
9. The information processing method according to any one of claims 1 to 8, wherein, In the presenting of the influence information, information indicating an inference engine from among the plurality of inference engines, which has outputted the object detection result related to the influence information to be presented, respectively, is also presented.
10. An information processing system, comprising: an inference result acquisition unit that acquires, for each of a plurality of inference engines, a plurality of object detection results for the same image as the inference engine; An input data influence obtaining unit that obtains, for each of the plurality of object detection results of each of the inference engines, influence information indicating whether or not the image has an influence on each of the plurality of object detection results, and, if so, whether or not the image has an influence on the plurality of object detection results, and whether or not the image has an influence on the plurality of object detection results; A determining unit configured to determine one or more combinations of object detection results of the same type among the plurality of object detection results of the respective inference engines based on types of detected objects of the plurality of object detection results of the respective inference engines, and And an influence presentation unit configured to present, via a presentation device, the influence information obtained for each of the plurality of object detection results of the inference engine, the object detection results included in the same combination, and the object detection results included in the same combination, the influence information being superimposed on each of the object detection results included in the same combination, the object detection results included in the same combination being superimposed on each other.
11. A computer-readable storage medium storing a program that, when executed by a processor in an information processing apparatus provided with the processor, causes the processor to: for each of a plurality of inference engines, obtaining a plurality of object detection results for the same image as the inference engine, Obtaining, for each of the plurality of object detection results of the respective inference engines, influence information indicating the presence or absence of an influence of the image on each of the plurality of object detection results and, if the influence is present, information indicating one or both of the magnitude and the directivity of the influence, Determining one or more combinations of object detection results of the same category among the plurality of object detection results of the respective inference engines based on the categories of the detected objects of the plurality of object detection results of the respective inference engines, The influence information acquired for each of the object detection results included in the same combination among the plurality of object detection results of the inference engine is superimposed on each of the object detection results included in the same combination via a presentation device, and then the object detection results included in the same combination, on which the influence information is superimposed, are presented in parallel with each other.

Description

Information processing method, information processing system, and storage medium Technical Field The present disclosure relates to an information processing method and the like executed by a computer. Background Techniques for presenting the result of detection of an object captured in an image have been proposed (for example, refer to patent document 1 and non-patent document 1). Prior art literature Patent literature Patent document 1 Japanese patent application laid-open No. 2018-181273 Non-patent literature Non-patent literature 1：Erik Bochinski et al,"High-Speed tracking-by-detection without using Image information",14th IEEE International Conference on Advanced Video and Signal Based Surveillance(AVSS),2017, 8 Disclosure of Invention Problems to be solved by the invention In the conventionally proposed technique, it is difficult to evaluate the behaviors of a plurality of different inference engines, each of which performs an inference process such as an object detection process, in parallel. The present disclosure provides an information processing method and the like capable of evaluating the behavior of each of a plurality of different inference engines in parallel. Means for solving the problems An information processing method according to an aspect of the present disclosure is a method executed by a computer, in which a plurality of inference results are obtained as a result of inference on the same input data by a plurality of inference engines, an influence of the input data on each of the plurality of inference results is obtained for each of the plurality of inference results, one or more combinations are determined based on the plurality of inference results, the one or more combinations are combinations of the plurality of inference results, and the influence obtained for each of the plurality of inference results included in the same combination is presented in parallel or overlapping with each other via a presentation device. The information processing system according to an aspect of the present disclosure includes an inference result acquisition unit that acquires a plurality of inference results that are results of inferences made on the same input data by a plurality of inference engines, an input data influence acquisition unit that acquires, for each of the plurality of inference results, an influence of the input data on each of the plurality of inference results, a determination unit that determines one or more combinations that are combinations of the plurality of inference results, based on the plurality of inference results, and an influence presentation unit that presents, via a presentation device, the influences acquired by the inference results included in the same combinations, respectively, in parallel or overlapping relation with each other. The modes of the summary and the concrete may be realized by a recording medium such as an apparatus, an integrated circuit, or a computer-readable CD-ROM, or by any combination of an apparatus, a system, an integrated circuit, a method, a computer program, and a recording medium, in addition to the above-described method and system. ADVANTAGEOUS EFFECTS OF INVENTION By the information processing method and the like of the present disclosure, the behavior of each of a plurality of different inference engines can be evaluated in parallel. Drawings Fig. 1 is a schematic diagram of a screen to which an information processing method of an embodiment is applicable. Fig. 2 is a schematic diagram for explaining differences in the results of respective object detections of different object detection models. Fig. 3A is a diagram schematically showing object detection results based on 3 object detection models for a common input image. Fig. 3B is a diagram showing an example of data of the object detection result set output from the object detection model. Fig. 4A is a diagram schematically showing an example of a state after the information processing method of the embodiment is applied to the object detection result shown in fig. 3A. Fig. 4B shows an example of data of the object detection result set after the information processing method of the embodiment is applied. Fig. 5A is a diagram for explaining the result of processing of inserting replacement data performed in the information processing method according to the embodiment. Fig. 5B is a diagram showing an example of data of an object detection result set corresponding to the result of the processing shown in fig. 5A. Fig. 6 is a diagram for explaining the insertion of substitute data by the information processing method according to the embodiment, and is different from the example shown in fig. 5A. Fig. 7A shows a display example in a UI screen to which the information processing method of the embodiment is applied. Fig. 7B shows a display example in a UI screen to which the information processing method of the embodiment is applied. Fig. 7C shows a display example in a UI screen