Abnormal Scene Detection via Motion Feature Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video surveillance systems lack the ability to determine and interpret the intent or actions of individuals in a video scene, such as distinguishing between a fight and other activities like conversations.
Innovation Solution
A system and method that identify at least two motion objects in a video scene, determine motion features such as movement trails, intensities, and consistencies, and assess whether they are involved in a fight by comparing these features against predefined criteria, using processors and data from video acquisition modules.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If video surveillance systems monitor human behavior, then security and monitoring capabilities are improved, but the ability to determine and interpret intent or actions of people remains insufficient
Solution Approach 1:
The system segments human behavior analysis into multiple independent components: motion detection, pose estimation, action recognition, and intent classification. Each component processes specific aspects of behavior separately before integrating results to determine overall intent, enabling comprehensive interpretation without requiring a single complex system
Solution Approach 2:
The system introduces intermediate processing layers including pose estimation models and action recognition algorithms that bridge the gap between raw video data and intent interpretation. These intermediaries extract meaningful features and transform them into interpretable action categories, enabling the system to infer intent from observable behaviors
2Measurement precision
If the system analyzes motion features of multiple objects to detect fights, then detection accuracy is improved, but computational complexity increases
Solution Approach 1:
The system divides fight detection into separate analytical stages: individual motion feature extraction, pairwise interaction analysis, and conflict pattern recognition. By processing motion features of multiple objects in segmented stages rather than simultaneously, the system achieves high detection accuracy while managing computational complexity through progressive refinement
Solution Approach 2:
The system applies different analysis depths to different regions of interest. When potential conflict is detected between specific object pairs, the system intensifies local analysis of their motion features while reducing analysis of other areas, thereby maintaining high detection precision for critical events without uniformly processing all objects at maximum complexity
Data Source
AI summary
A method for detecting abnormal scene may include obtaining data relating to a video scene, identify at least two motion objects in the video scene based on the data and determining a first motion feature relating to the at least two motion objects based on the data. The method may also include determining a second motion feature relating to at least one portion of each of the at least two motion objects based on the data. The method may further include determining whether the at least two motion objects are involved in a fight based on the first motion feature and the second motion feature.