Object Recognition System Frame Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing facial recognition technologies face high calculation loading and unstable recognition results due to the need to process multiple faces in image frames and environmental interruptions, especially when dealing with continuous video streams.
Innovation Solution
An object recognition method and system that detects and tracks objects across frames, determining if they represent the same target object based on position and confidence level, allowing for the assignment of recognition results rather than repeated calculations, thereby reducing processing load and improving stability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If object recognition is performed on every frame of video stream, then recognition accuracy is maintained, but calculation loading increases significantly
Solution Approach 1:
The patent applies local quality by performing full object recognition only on specific frames (when objects are first detected or when confidence levels need verification) rather than uniformly processing every frame. This selective approach maintains recognition accuracy for critical moments while reducing calculation loading during continuous tracking phases where objects are already identified.
2Productivity
If object recognition is performed on every frame of video stream, then recognition results are obtained continuously, but processing time increases
Solution Approach 1:
The patent implements preliminary action by performing object detection and initial recognition on key frames first, then using the results from these preliminary processing steps to inform subsequent frame processing decisions. This allows the system to maintain recognition continuity while reducing overall processing time by avoiding redundant full recognition operations on every frame.
3Adaptability or versatility
If recognition is performed on multiple faces in a single image frame, then comprehensive identification is achieved, but calculation loading increases
Solution Approach 1:
The patent applies segmentation by dividing the image frame into multiple regions of interest, each containing a detected face or object. The system processes each segment independently for recognition, which maintains the ability to identify multiple faces simultaneously while reducing overall calculation loading by avoiding redundant processing of the entire image and allowing parallel processing of individual face segments.
Data Source
AI summary
An object recognition method and system thereof are provided. A recognition result of a first object of a (i−1)th frame of a video stream is obtained. A ith frame is received, and a second object is detected from the ith frame. Whether the first object and the second object are corresponding to the same target object is determined according to a position of the first object in the (i−1)th frame and a position of the second object in the ith frame. If the first object and the second object are corresponding to the same target object, whether a recognition confidence level is greater than a predetermined threshold is determined so as to perform the object recognition on the second object or assign the recognition result of the first object to the second object.