US-20260127887-A1 - INFORMATION PROCESSING SYSTEM, INFORMATION PROCESSING APPARATUS, INFORMATION PROCESSING METHOD, AND NON-TRANSITORY COMPUTER-READABLE MEDIUM

US20260127887A1US 20260127887 A1US20260127887 A1US 20260127887A1US-20260127887-A1

Abstract

An information processing system ( 100 ) includes an analysis result acquisition unit ( 131 ), a candidate detection unit ( 132 ), and a stray detection unit ( 134 ). The analysis result acquisition unit ( 131 ) acquires an analysis result of videos captured by a plurality of image capture apparatuses ( 101 ). The candidate detection unit ( 132 ) detects a stray candidate from persons captured in the videos by using person attributes included in the analysis result and a candidate condition. In a case where a companion of the stray candidate exists at a first point in time, the stray detection unit ( 134 ) detects a stray from among the stray candidates, based on a comparison result between a companion of the stray candidate at the first point in time and a companion of the stray candidate at a second point in time earlier than the first point in time.

Inventors

Takayuki KASE
Jianquan Liu
Tingting DONG
Noboru Yoshida

Assignees

NEC CORPORATION

Dates

Publication Date: 20260507
Application Date: 20221011

Claims (19)

1 . An information processing system comprising: at least one memory storing instructions; and at least one processor configured to: acquire an analysis result of videos captured by a plurality of image capture units; detect a stray candidate from one or more persons captured in the video by using one or more person attributes included in the analysis result and a candidate condition; and in a case where a companion of the stray candidate exists at a first point in time, detect a stray from among the one or more stray candidates, based on a comparison result between a companion of the stray candidate at the first point in time and a companion of the stray candidate at a second point in time earlier than the first point in time.
2 . The information processing system according to claim 1 , wherein the candidate condition includes a condition related to age.
3 . The information processing system according to claim 1 , wherein the at least one processor is further configured to acquire feature information of a stray being a detection target, wherein the candidate condition includes feature information of the stray.
4 . The information processing system according to claim 1 , wherein the at least one processor is configured to, in a case where a companion of the stray candidate exists at a first point in time, detect a stray from among the one or more stray candidates, based on whether a companion of the stray candidate has changed between the first point in time and the second point in time.
5 . The information processing system according to claim 1 , wherein the at least one processor is configured to, in a case where a companion of the stray candidate exists at a first point in time, detect a stray from among the one or more stray candidates, based on the comparison result and a degree of danger based on a position of the stray candidate at a first point in time.
6 . The information processing system according to claim 1 , the at least one processor is further configured to determine a group to which a person in the video belongs by using a person attribute included in the analysis result and a grouping condition for grouping one or more persons captured in the video; in a case where a companion of the stray candidate exists at a first point in time, compare a companion of the stray candidate between the first point in time and the second point in time by using groups to which the stray candidate belongs at the first point in time and the second point in time, respectively; and detect a stray from among the one or more stray candidates, based on the comparison result.
7 . The information processing system according to claim 6 , wherein the at least one processor is further configured to: identify whether a companion of the stray candidate exists at the first point in time by using a group to which the stray candidate belongs at the first point in time; and in a case where a companion of the stray candidate is identified to exist at the first point in time, compare a person belonging to a same group as the stray candidate between the first point in time and the second point in time and detects a stray from among the one or more stray candidates, based on the comparison result.
8 . The information processing system according to claim 6 , wherein the at least one processor is further configured to acquire feature information of a stray being a detection target; and determine a group to which a person in the video belongs by further using feature information of the stray.
9 . The information processing system according to claim 1 , wherein the at least one processor is further configured to predict a moving range of a person captured in the video by using the person attribute.
10 . The information processing system according to claim 9 , wherein the at least one processor is configured to predict a moving range of the person by further using layout information about a location images of which are captured by the plurality of image capture units.
11 . The information processing system according to claim 9 , wherein the at least one processor is further configured to: detect a moving pattern of a person captured in the video, based on a person attribute between the first point in time and the second point in time, and predict a moving range of a person captured in the video between the first point in time and the second point in time by further using the moving pattern.
12 . The information processing system according to claim 11 , wherein the at least one processor is configured to: set a moving range predicted for the stray candidate as a search range of the stray candidate; and detect a stray candidate from one or more persons captured in the search range.
13 . The information processing system according to claim 9 , wherein the at least one processor is further configured to: cause a display unit to display stray information related to the detected stray; and predict a moving range of the detected stray, and the stray information includes the predicted moving range.
14 . The information processing system according to claim 13 , wherein the stray information further includes the layout information, and the at least one processor is configured to cause the display unit to display an image acquired by superposing the predicted moving range on the layout information.
15 . The information processing system according to claim 13 , wherein the stray information includes at least one of an image of the detected stray and a position of the detected stray at the first point in time.
16 . The information processing system according to claim 13 , wherein the at least one processor is configured to, in a case where there are a plurality of the detected strays, cause the display unit to display pieces of the stray information of the plurality of strays in descending order of degree of danger at the first point in time.
17 . (canceled)
18 . An information processing method comprising, by one or more computers: acquiring an analysis result of videos captured by a plurality of image capture units; detecting a stray candidate from one or more persons captured in the video by using one or more person attributes included in the analysis result and a candidate condition; and, in a case where a companion of the stray candidate exists at a first point in time, detecting a stray from among the one or more stray candidates, based on a comparison result between a companion of the stray candidate at the first point in time and a companion of the stray candidate at a second point in time earlier than the first point in time.
19 . A non-transitory computer-readable medium storing an information processing program that causes one or more computers to execute: acquiring an analysis result of videos captured by a plurality of image capture units; detecting a stray candidate from one or more persons captured in the video by using one or more person attributes included in the analysis result and a candidate condition; and, in a case where a companion of the stray candidate exists at a first point in time, detecting a stray from among the one or more stray candidates, based on a comparison result between a companion of the stray candidate at the first point in time and a companion of the stray candidate at a second point in time earlier than the first point in time.

Description

TECHNICAL FIELD The present invention relates to an information processing system, an information processing apparatus, an information processing method, and a medium. BACKGROUND ART For example, Patent Document 1 discloses a technology for detecting a stray. A stray determination unit described in Patent Document 1 particularly extracts only a person of an age to become a possible stray, based on person information. The person information is a result of performing feature extraction of a contour and/or the like by a feature extraction unit, a person extraction unit, and a personal feature analysis unit from an image of a surveillance camera installed at a certain location, automatically recognizing persons, and performing personal feature analysis on the age, the clothing, the physical constitution, and the like of each person. In a case where the stray determination unit described in Patent Document 1 determines that a person may be a stray, based on information such as an anxious look or action and whether the person is acting independently (alone), the information being a result of an action analysis on the person performed by an action analysis unit in parallel, the stray determination unit determines that such a person is a stray. Note that Patent Document 2 describes a technology for computing a feature value of each of a plurality of keypoints of a human body included in an image, searching for an image including a human body with a similar pose or a human body with a similar movement, based on the computed feature values, and grouping together and categorizing the human bodies with similar poses and similar movements. Non-Patent Document 1 describes a technology related to skeletal estimation of a person. RELATED DOCUMENT Patent Document Patent Document 1: Japanese Patent Application Publication No. 2021-108149 Patent Document 2: International Application Publication No. WO 2021/084677 Non-Patent Document Non-Patent Document 1: Zhe Cao, Tomas Simon, Shih-En Wei, Yaser Sheikh, “Realtime Multi-Person 2D Pose Estimation using Part Affinity Fields,” The IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2017, PP. 7291 to 7299 SUMMARY Technical Problem However, as described above, the technology described in Patent Document 1 detects a stray, based on information such as an anxious look or action and whether a person is acting independently (alone). Therefore, high-precision detection of a stray abducted by a strange third person is difficult. For example, high-precision detection of a look or an action of a person in an image from the image is generally difficult. Even in a case where detection is possible, high-precision detection of a look or an action of the person in the image may not be possible in a case of poor image quality or the like. Thus, in a case where detection precision of an anxious look or action is low, the technology described in Patent Document 1 may not be able to perform high-precision detection of an abducted stray. Further, for example, a person of an age to become a possible stray may generally have an anxious look or take an anxious action even in a case of being with a protector. In such a case, the person may be detected as a stray even in the case of being with a protector in the technology described in Patent Document 1. Further, for example, an abducted stray is likely to act with a strange third person and is not likely to act independently. Therefore, it is difficult to detect an abducted stray by using an independent (alone) action in the technology described in Patent Document 1. Abduction is highly likely a dangerous situation for a target stray and detection of the abduction is extremely important for safety and the like of the stray. Note that Patent Document 2 and Non-Patent Document 1 do not disclose a technology for detecting a stray. An example of an object of the present invention is to, in view of the aforementioned issues, provide an information processing system, an information processing apparatus, an information processing method, and a medium that resolve the issue of ensuring safety of a stray. Solution to Problem According to an aspect of the present invention, an information processing system including: an analysis result acquisition unit that acquires an analysis result of videos captured by a plurality of image capture units;a candidate detection unit that detects a stray candidate from one or more persons captured in the video by using one or more person attributes included in the analysis result and a candidate condition; anda stray detection unit that, in a case where a companion of the stray candidate exists at a first point in time, detects a stray from among the one or more stray candidates, based on a comparison result between a companion of the stray candidate at the first point in time and a companion of the stray candidate at a second point in time earlier than the first point in time is provided. According to an aspect of