Video Scene Analysis for Hidden Associate Identification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video surveillance systems are inadequate in identifying potential associates of organized crime group members, especially when they avoid direct communication and are not captured together on camera, making it difficult for law enforcement to monitor and discover their networks.
Innovation Solution
A method and device that analyze video surveillance footage by establishing video scenes extending before and after a target person's appearance, identifying individuals who appear in multiple scenes as potential associates, even if they are not co-appearing together, by setting predetermined durations for each scene.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If video surveillance systems only capture direct co-appearances of target persons and associates, then the system maintains simple analysis criteria, but it fails to identify hidden associates who avoid direct contact
Solution Approach 1:
The system performs preliminary actions by establishing video scenes that extend before and after the target person's appearance. This allows the system to capture associates who may appear before the target arrives or after the target leaves, even if they never appear simultaneously with the target in the frame. The predetermined duration extensions enable the system to proactively include potential associate appearances in the analysis scope.
Solution Approach 2:
The system transitions from analyzing only spatial co-occurrence (whether persons appear together in the same video frame) to analyzing temporal relationships (appearances within extended time windows before and after target appearances). This dimensional shift from spatial to temporal analysis enables the system to identify associates through temporal proximity patterns rather than requiring simultaneous visual presence.
2Measurement precision
If the system extends video scene duration to capture more potential associates, then the probability of identifying associates increases, but the amount of video data to be processed increases
Solution Approach 1:
By pre-defining the scene extension duration before target appearance and post-appearance, the system establishes fixed temporal boundaries for analysis. This preliminary action allows the system to process only the necessary extended time windows rather than analyzing entire video streams, making the increased data processing manageable through predetermined parameters.
Solution Approach 2:
The system uses parameter changes by adjusting the predetermined duration values to control the extent of scene extension. By modifying these time parameters, the system can balance between capturing sufficient associate appearance data and limiting the total video data volume to processable levels. The threshold number of scenes parameter also provides additional control over the analysis scope.
3Measurement precision
If the system analyzes multiple video scenes with extended durations, then hidden associates can be discovered through pattern recognition, but the processing time and computational resources increase
Solution Approach 1:
The system applies local quality by focusing analysis resources on specific temporal regions of interest - the extended windows before and after target appearances. Rather than uniformly analyzing all video content, the system concentrates processing on these locally defined scenes where associate appearances are most likely to occur, improving efficiency by directing computational resources to high-probability areas.
Solution Approach 2:
The system employs parameter changes by using a threshold number of scenes to filter results. By requiring associates to appear in more than a threshold number of extended scenes, the system can adjust its sensitivity and processing scope. This parameter allows tuning between comprehensive analysis (lower threshold, more processing) and efficient analysis (higher threshold, less processing) based on operational requirements.
Data Source
AI summary
There is provided a method for identifying potential associates of at least one target person, the method comprising: providing a plurality of videos; identifying appearances of the at least one target person in the plurality of videos; establishing a plurality of video scenes from the plurality of videos, wherein each one of the plurality of video scenes begins at a first predetermined duration before a first appearance of the at least one target person in the respective video scene and ends at a second predetermined duration after a last appearance of said at least one target person in the respective video scene; determining individuals who appear in more than a predetermined threshold number of the plurality of video scenes; and identifying the individuals as potential associates of the at least one target person.