Object Tracking Search Range Update to Prevent Passing Over
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing object tracking technologies suffer from the phenomenon of 'passing over', where a similar object is mistakenly tracked instead of the target, making it difficult to return to the correct target once the error occurs.
Innovation Solution
An object tracking device that extracts target candidates from time series images, updates a search range based on frame and movement information, tracks using a confidence level with a target model, and updates the model using these candidates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the search range is expanded to cover the entire image to ensure the target is found, then the probability of finding the target increases, but the computational complexity and processing time increase significantly
Solution Approach 1:
The patent applies local quality by focusing computational resources on a specific local region (search range) rather than the entire image. The search range is dynamically adjusted based on target movement information, concentrating processing power where the target is likely to be located, thus maintaining high tracking accuracy while reducing overall computational complexity.
Solution Approach 2:
The patent uses preliminary action by predicting the target's future position based on movement information from previous frames before actually performing the full search. This preliminary estimation allows the system to narrow down the search range in advance, reducing the computational burden of the subsequent tracking operation.
2Productivity
If the search range is reduced to a small area to decrease computational load, then processing speed increases, but the risk of missing the target or similar objects increases
Solution Approach 1:
The patent applies dynamics by making the search range adaptive and dynamic rather than fixed. The search range automatically adjusts its size and position based on the target's movement characteristics and prediction accuracy, ensuring it remains appropriately sized to capture the target while minimizing unnecessary computational areas.
3Adaptability or versatility
If the target model is continuously updated using extracted candidates to improve tracking adaptability, then the system can adapt to appearance changes, but the risk of learning wrong features from similar objects increases
Solution Approach 1:
The patent applies feedback by using the tracking results and confidence levels to control model updates. The target model is updated based on feedback from successful tracking instances where the confidence level indicates high similarity, while avoiding updates from potential false positives. This feedback mechanism allows the system to adapt to appearance changes while maintaining feature discrimination accuracy.
Data Source
AI summary
In an object tracking device, an extraction means extracts target candidates from time series images. A search range update means updates a search range based on frame information of a target in a previous image in a time series and movement information of the target. A tracking means searches for and tracks the target using a confidence level indicating similarity with a target model, from among the target candidates extracted in the search range. A model update means updates the target model using the target candidates extracted in the search range.