Target Tracking via Candidate Position Combination
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional visual target tracking systems face challenges in accurately tracking targets with fast motion or partial occlusion, leading to blurred images and incorrect tracking results.
Innovation Solution
The method employs a region proposal network and feature maps to generate candidate positions of a target, combining multiple anchor boxes' probabilities and deviations to improve tracking robustness and precision, effectively addressing issues of occlusion and fast motion.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional target tracking systems use single position detection, then the system complexity is low, but the tracking precision deteriorates when target is occluded or moving fast
Solution Approach 1:
The patent divides the target detection into multiple candidate positions (at least two) instead of relying on a single detection result. Each candidate position is evaluated independently, and the system segments the detection task into multiple proposals that are then combined to determine the final target position, improving precision without requiring complex hardware changes
Solution Approach 2:
The patent combines multiple candidate position detections to determine the final target position. By merging the results from multiple detection attempts and using combination strategies (such as voting or probability integration), the system achieves more reliable tracking results under challenging conditions like occlusion and fast motion
2Speed
If the system tracks fast moving targets, then the tracking speed is high, but the image quality deteriorates causing motion blurring
Solution Approach 1:
The system performs preliminary detection at multiple candidate positions before final target confirmation. By anticipating possible target positions based on motion patterns and generating multiple proposals in advance, the system prepares for motion blur scenarios without sacrificing tracking speed, as the multiple detections are processed in parallel or near-real-time
3Reliability
If the system handles occluded targets, then the tracking robustness should improve, but the detection difficulty increases
Solution Approach 1:
The system performs more detection actions than strictly necessary by generating at least two candidate positions for each target. This excessive detection approach ensures that even when part of the target is occluded, sufficient candidate positions remain available for accurate tracking, as the system doesn't require complete target visibility to function effectively
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The present disclosure discloses a method and apparatus for tracking a target, a device, a medium and a computer program product, and relates to the field of computer vision technology. A specific embodiment includes: generating a position of a candidate box of a to-be-tracked target in a to-be-processed image; determining, for a pixel in the to-be-processed image, a probability that each anchor box of at least one anchor box arranged for the pixel includes the to-be-tracked target, and determining a deviation of the candidate box corresponding to the anchor box relative to the anchor box; determining candidate positions of the to-be-tracked target corresponding to the at least two anchor boxes respectively; and combining at least two candidate positions among the determined candidate positions to obtain a position of the to-be-tracked target in the to-be-processed image. The present disclosure can select at least two candidate positions of a to-be-tracked target, and combine the candidate positions, thereby effectively avoiding the problem that the target is difficult to track because the target is blurred due to the target being occluded or moving fast, and improving the robustness and precision of the tracking system.