Occlusion Detection in Video-Based Object Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional video surveillance systems require manual review of large amounts of data, hindering real-time decision-making due to the challenge of accurately detecting occlusions in video-based object tracking, which can lead to inaccurate object representation and identity switches.
Innovation Solution
A process and device for occlusion detection in video-based object tracking that computes histogram and depth level data for video frames, comparing variations to threshold values to determine occlusions, and generates output data indicating occluded frames, utilizing Region of Interest (ROI) analysis and depth level estimation to filter out false positives.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual review of surveillance video data is used, then detection accuracy can be maintained, but productivity is reduced due to the large amount of data requiring review
Solution Approach 1:
The patent introduces an automated occlusion detection system that acts as an intermediary between raw video data and human operators. The system computes histogram data and depth level data from video frames, compares variations against threshold values, and automatically identifies occluded frames, thereby reducing the burden on manual reviewers while maintaining detection accuracy
Solution Approach 2:
The patent replaces the mechanical manual review process with an automated computational system. The system uses computer vision algorithms to compute histogram data representing color distributions and depth level data representing spatial relationships, then automatically compares these features to detect occlusions, substituting human visual inspection with machine-based analysis
2Productivity
If automated occlusion detection is implemented, then productivity is improved, but measurement precision may worsen due to false positives from inaccurate occlusion detection
Solution Approach 1:
The patent moves occlusion detection from a single-dimensional approach (relying solely on histogram color distribution changes) to a multi-dimensional approach by integrating both histogram data and depth level data. The depth level information provides spatial context about object positions and occlusion relationships, enabling more accurate detection and reducing false positives while maintaining high processing throughput
3Device complexity
If histogram analysis alone is used for occlusion detection, then device complexity is reduced, but measurement precision worsens due to inability to distinguish occlusions from other changes
Solution Approach 1:
The patent merges two complementary detection approaches: histogram-based color distribution analysis and depth-level-based spatial relationship analysis. By combining these two methods, the system achieves more accurate occlusion detection than either method alone, as histogram changes may result from various factors while depth level changes provide specific information about occlusion geometry and object positioning
Data Source
AI summary
Processes, systems, and devices for occlusion detection for video-based object tracking (VBOT) are described herein. Embodiments process video frames to compute histogram data and depth level data for the object to detect a subset of video frames for occlusion events and generate output data that identifies each video frame of the subset of video frames for the occlusion events. Threshold measurement values are used to attempt to reduce or eliminate false positives to increase processing efficiency.