Target Tracking With Progressive Networks for Tracking Loss Recovery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional target tracking methods struggle with maintaining accuracy when targets are deformed, move, or become partially occluded, leading to a high likelihood of tracking loss.
Innovation Solution
A target tracking method utilizing a progressive network architecture with shallow, low-resolution networks followed by deep, high-resolution networks to adaptively adjust search depth and image resolution based on tracking results, ensuring accurate and efficient target detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a single deep network with high resolution is used for target tracking, then tracking accuracy is improved, but power consumption and algorithm complexity increase significantly
Solution Approach 1:
The patent divides the target tracking task into two segments: a first network performs initial tracking in normal scenarios, and a second network performs detailed search only when the target is lost. This segmentation allows the system to use simple algorithms most of the time while reserving complex algorithms for specific needs, thereby reducing overall algorithm complexity while maintaining tracking accuracy.
Solution Approach 2:
The patent applies different network configurations to different scenarios: the first network uses lower computational resources for normal tracking conditions, while the second network employs higher resolution and deeper architecture only when target loss is detected. This local quality approach ensures high tracking accuracy is achieved only where necessary, reducing overall system complexity.
2Measurement precision
If a single deep network with high resolution is used for target tracking, then tracking accuracy is improved, but power consumption increases
Solution Approach 1:
The patent segments the tracking process into two stages with different power requirements. The first network consumes less power for routine tracking, and the second network consumes more power only when target loss occurs. This segmentation ensures high power consumption is avoided during normal operation while maintaining tracking accuracy when needed.
Solution Approach 2:
The system applies high-power deep network processing only locally to specific frames where target loss is detected, rather than continuously using high-power processing for all frames. This local quality approach maintains tracking accuracy during critical moments while minimizing overall power consumption during normal tracking operations.
3Ease of operation
If conventional target tracking methods are used, then the system is simple to operate, but tracking accuracy deteriorates when targets are deformed, move, or occluded
Solution Approach 1:
The patent implements a dynamic tracking system that automatically switches between the first network and second network based on tracking status. When the target is successfully tracked, the simpler first network is used; when target loss is detected, the system dynamically transitions to the more capable second network. This dynamic approach maintains system simplicity in normal operation while achieving high tracking accuracy in challenging scenarios.
Solution Approach 2:
The system incorporates feedback mechanisms where the tracking result of the first network determines whether to activate the second network. If the first network successfully tracks the target, no further action is needed; if tracking is lost, the feedback triggers the second network to perform a detailed search. This feedback-based approach maintains operational simplicity while improving tracking accuracy when needed.
Data Source
AI summary
A target tracking method and an electronic device. The method includes, when a first region displays an nth frame of a video stream, receiving a first operation input to the nth frame of image by a user; in response to the first operation, performing feature extraction on the nth frame of image, to obtain a template feature, where the nth frame of image includes a target; determining, through a first network based on the template feature, whether an (n+x)th frame of image of the video stream includes the target; and if it is determined, through the first network, that the (n+x)th frame of image does not include the target, determining, through a second network based on the template feature, whether the (n+x)th frame of image includes the target, where the first network and the second network satisfy at least one condition in a first preset condition.