Occluded Object Detection via Fingerprint Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Multiple object detection systems (MOTS) face challenges in accurately tracking objects when one object is occluded, leading to detection inaccuracies and reduced efficiency due to reliance on limited information and failure to recognize occluded objects.
Innovation Solution
The system enhances occluded object detection by receiving video frames, generating object identifier mapping information, detecting multi-object tracking inaccuracies, and updating identifiers based on fingerprint comparisons between current and historic frames, thereby improving tracking accuracy without reducing efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If object detection systems rely on limited information from video frames, then the system complexity remains low, but the detection accuracy of occluded objects deteriorates
Solution Approach 1:
The system performs preliminary actions by detecting objects in previous video frames before the current frame, storing their identifiers and spatial information. When an object is occluded in the current frame, the system uses this pre-detected information to maintain tracking continuity, thereby improving detection accuracy without requiring complex real-time processing during occlusion events.
Solution Approach 2:
The system introduces an intermediary mechanism (object identifier and spatial information storage) that mediates between detected objects across different video frames. This intermediary allows the system to bridge gaps caused by occlusions by matching objects based on stored identifiers and spatial relationships rather than relying solely on direct visual detection in each frame.
2Reliability
If the system tracks all objects using detailed information, then tracking accuracy improves, but the processing time and computational load increase
Solution Approach 1:
The system extracts only the essential and most useful information from detected objects, specifically object identifiers and spatial information (position, size). By taking out only these critical parameters for storage and comparison, the system maintains reliable tracking accuracy while minimizing computational load and processing time, avoiding the need to process and store all detailed object characteristics.
3Productivity
If the system uses simple object identification methods, then the processing speed remains high, but the ability to distinguish occluded objects from new objects deteriorates
Solution Approach 1:
The system implements a universal object identification approach that serves multiple functions: it distinguishes occluded objects from new objects, maintains tracking continuity, and enables efficient processing. By using consistent object identifiers and spatial information comparison across all video frames, the system achieves both high processing speed and accurate object distinction without requiring separate specialized methods for different scenarios.
Data Source
AI summary
A system may be configured to perform enhanced detection of occluded objects in a multiple object detection system. In some aspects, the system may detect a plurality of current objects in a current video frame, generate initial object identifier mapping information assigning a candidate object identifier to a first object of the plurality of current objects, and detect a multi-object tracking (MOT) inaccuracy condition. Further, the system may generate a current fingerprint corresponding to the first object of the plurality of current objects, identify an assignment error based on comparing the current fingerprint to historic fingerprints associated with a plurality of historic objects detected in a previous frame, generate updated object identifier mapping information based on the assignment error, and display the current video frame based on the updated object identifier mapping information.