Tracking System Dynamic Mode Switching for Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing person tracking systems face reduced recognition rates for distant faces and misjudgment of interlaced, background, or re-entering faces, leading to low tracking accuracy, and the parallel execution of face recognition and tracking programs increases computational load, making practical implementation difficult.
Innovation Solution
A tracking system and method that synergistically operate face recognition and tracking programs by using an image capture device, memory, and processor to capture and process video streams, identify faces, and determine movement based on predetermined distances, allowing efficient switching between recognition and tracking modes to reduce power consumption and improve overall system efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If face recognition program and face tracking program are executed in parallel, then tracking coverage is improved, but computational load increases making practical implementation difficult
Solution Approach 1:
The system dynamically switches between face recognition program and human figure tracking program based on detection conditions. When a face is detected within the third threshold distance, the face recognition program is executed; otherwise, the human figure tracking program is executed. This dynamic switching resolves the contradiction by adapting the computational approach to the current situation, maintaining tracking accuracy while avoiding the excessive computational load of parallel execution.
2Measurement precision
If face recognition program is used for distant faces, then recognition rate is maintained, but system complexity increases
Solution Approach 1:
The system applies different recognition approaches based on the local condition of face distance. For faces within the third threshold distance, face recognition program is used with high precision; for faces beyond this distance, human figure tracking program is used. This local quality principle resolves the contradiction by matching the recognition method to the specific distance condition, maintaining recognition rate for distant faces without unnecessarily increasing system complexity for all cases.
3Stability of the object's composition
If face tracking program is used for all distances, then tracking continuity is improved, but recognition accuracy for distant faces deteriorates
Solution Approach 1:
The system dynamically adjusts the tracking program based on the detected face distance. When the face is within the third threshold distance, face recognition program is executed for high accuracy; when the face is beyond this distance, human figure tracking program is executed to maintain tracking continuity. This dynamic adaptation resolves the contradiction by optimizing the tracking approach according to the specific distance condition.
Solution Approach 2:
Different tracking programs are applied to different distance zones. The face recognition program handles close-range tracking with high precision, while the human figure tracking program handles distance-range tracking for continuity. This local quality approach ensures that each distance zone receives the appropriate tracking method, resolving the contradiction between continuity and accuracy.
Data Source
AI summary
A tracking system includes an image capture device, a memory and a processor. The image capture device is used to capture a video stream in a target area. The memory stores face information. The processor reads the face information from the memory and reads the video stream from the image capture device. The processor is used to identify a face matching the face information in the video stream to track the movement of the face in the target area. The processor is used to determine whether a distance between the face and the image capture device is greater than a predetermined distance. The processor is used to track the movement of a human figure corresponding to the face in the target area when the processor determines that the distance is greater than the predetermined distance.