Robot Target Tracking Using Cached Features to Reduce False Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current robots with human face tracking functions suffer from high false recognition rates, leading to poor tracking performance.
Innovation Solution
A target tracking method that determines and stores a target object feature, matches recognized object features with the stored feature, and re-finds the target object by matching with cached features upon loss, reducing interference and false recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If object feature recognition is performed continuously to improve tracking reliability, then false recognition rate increases, but tracking reliability deteriorates
Solution Approach 1:
The system performs preliminary action by caching the target object feature in advance before tracking begins. This cached feature serves as a reference template that enables rapid comparison and identification during tracking, eliminating the need for continuous complex recognition operations and reducing false recognition while maintaining reliability
Solution Approach 2:
The system creates a copy of the target object feature and stores it in cache memory. This copied feature template is then used for quick comparison against recognized objects during tracking, significantly reducing computational complexity and false recognition rates while maintaining high tracking reliability
2Measurement precision
If complex feature matching is performed to reduce false recognition, then processing time increases, but productivity decreases
Solution Approach 1:
The target object feature is extracted and cached in advance before tracking operations begin. This preliminary extraction creates a ready-to-use reference template that eliminates the need for repeated complex feature extraction during tracking, significantly improving processing speed while maintaining accurate matching through the pre-established template
Solution Approach 2:
The system extracts only the essential target object feature and stores it separately in cache memory, isolating this critical information from the main processing flow. This extraction enables rapid comparison operations during tracking without requiring complex real-time analysis, thus improving processing speed while maintaining low false recognition rates
Data Source
AI summary
The present application provides a target tracking method, a target tracking apparatus, and a robot. The target tracking method includes: determining a target object to be tracked, and storing a target object feature of the target object; in response to an object feature of a recognized object matching the target object feature that is pre-stored, tracking the recognized object; and in response to the object feature of the recognized object not matching the target object feature that is pre-stored, proceeding with object feature recognition of an object within a field of view; and in a case where the object feature matching the target object feature that is pre-stored is recognized again, tracking the object that is recognized once again.