Face Location Tracking via Predicted Region Similarity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing face tracking methods in image processing fail to maintain continuous tracking in complex scenarios such as sudden light changes, light interference, or rapid face movement, leading to blurred or discontinuous face images and compromised user experience, especially in devices with limited processing capabilities.
Innovation Solution
A face location tracking method that predicts a face region in the next image frame based on the current frame, searches for a preselected region with high similarity, and uses a predetermined selection rule to determine the face location tracking result, ensuring continuous tracking even in complex scenarios.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If face detection is performed on each image frame separately, then face location can be obtained in each frame, but face tracking loss or detection error occurs in complex scenarios such as sudden light change, light interference, or rapid face movement
Solution Approach 1:
The patent performs face detection in advance on a reference image frame to obtain a face region, then uses this pre-detected face region to predict face locations in subsequent image frames. This preliminary detection action allows the system to maintain continuous tracking in complex scenarios by relying on the pre-obtained face information rather than detecting each frame independently.
Solution Approach 2:
The patent introduces a predicted face region as an intermediary between the detected face region in the reference frame and the actual face regions in subsequent frames. This predicted region serves as a mediator that guides the search and selection process, enabling continuous tracking by bridging the gap between discrete detection results.
2Device complexity
If face detection is performed on each image frame independently, then processing is simpler, but tracking results become discontinuous and user experience deteriorates
Solution Approach 1:
By performing face detection once on a reference image frame in advance, the patent avoids the need to independently detect each subsequent frame, thereby reducing overall processing complexity while maintaining tracking continuity through the use of predicted regions and selection rules.
Solution Approach 2:
The patent establishes continuous tracking by using the detected face region from the reference frame to generate predicted face regions across multiple subsequent frames, ensuring that the useful action of face tracking continues without interruption even when individual frame detections fail.
3Use of energy by moving object
If traditional face tracking methods are used in terminal devices with limited processing performance, then resource consumption is lower, but face tracking accuracy and smoothness are insufficient
Solution Approach 1:
The patent extracts only the necessary face region information from the reference image frame and uses this extracted information to predict face locations in subsequent frames, rather than processing entire frames. This extraction approach reduces resource consumption while maintaining tracking accuracy.
Solution Approach 2:
The patent creates a predicted face region by copying and transforming the detected face region from the reference frame, then uses this copied region to guide the selection of actual face regions in subsequent frames. This copying mechanism enables accurate tracking with minimal processing resources.
Data Source
AI summary
A first face region within a first image is determined. The first face region includes a location of a face within the first image. Based on the determined first face region within the first image, a predicted face region within a second image is determined. A first region of similarity within the predicted face region is determined. The first region of similarity has at least a predetermined degree of similarity to the first face region within the first image. Whether a second face region is present within the second image is determined. The location of the face within the second image is determined based on the first region of similarity, the determination of whether the second face region is present within the second image, and a face region selection rule.