Video Movement-Route Discrepancy Analysis for Suspicious-Person Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies struggle to accurately identify specific individuals, such as suspicious persons, from camera images due to diverse behavioral patterns and the difficulty in preparing training data for machine learning models.
Innovation Solution
An information processing program that predicts movement routes of individuals in video data, using a machine learning model to analyze the difference between predicted and actual movement routes, allowing for the identification of individuals influenced by specific events, such as the presence of a police officer or advertisement, and specifying them as suspicious persons.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If machine learning models are trained using conventional methods with diverse behavioral patterns, then the ability to detect suspicious persons is improved, but the difficulty in preparing training data increases
Solution Approach 1:
The system performs preliminary actions by automatically extracting movement routes and predicting future positions before actual suspicious behavior occurs. The machine learning model is trained on predicted movement routes rather than requiring complex manual annotation of diverse behavioral patterns, thereby simplifying training data preparation while maintaining detection accuracy.
Solution Approach 2:
The system uses self-service by automatically generating training data through movement route extraction and prediction. The machine learning model processes video data autonomously to create labeled movement route datasets, eliminating the need for manual annotation and reducing the complexity of training data preparation while improving detection reliability.
2Reliability
If operators visually check video to detect suspicious persons, then detection accuracy is maintained, but the burden on operators increases
Solution Approach 1:
The system replaces the mechanical manual inspection process with an automated machine learning-based detection system. The machine learning model processes video data automatically to identify suspicious movement patterns, substituting operator visual inspection with computational analysis, thereby reducing operator burden while maintaining detection accuracy.
Solution Approach 2:
The system implements self-service by enabling the machine learning model to autonomously detect suspicious persons without human intervention. The model automatically analyzes movement routes, predicts future positions, and identifies suspicious behavior patterns, thereby eliminating the need for continuous operator monitoring and reducing operational burden.
3Reliability
If machine learning models are trained with comprehensive behavior patterns, then detection capability is improved, but the time required for training increases
Solution Approach 1:
The system performs preliminary extraction of movement routes from video data and generates predicted movement routes automatically. This preliminary processing creates ready-to-use training data that reduces the time required for model training while maintaining comprehensive detection capability through automated data generation.
Solution Approach 2:
The system substitutes manual training data preparation with automated machine learning processing. The machine learning model automatically extracts, processes, and generates training data from video sequences, dramatically reducing training time while maintaining comprehensive detection capability through systematic automated analysis.
Data Source
AI summary
A non-transitory computer-readable recording medium has stored therein an information processing program that causes a computer to execute a process comprising acquiring a video when a specific event has occurred specifying a first movement route of a person in a first period contained in the acquired video predicting a second movement route of the person in a second period after the first period based on the first movement route specifying an actual third movement route of the person in the second period by analyzing the acquired video and specifying a person related to the specific event from the video based on the second movement route and the third movement route.