Face Tracking via Touch-Triggered ROI Extraction in Video Calls
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video calling technologies lack advanced face tracking features, limiting user interaction and video quality during video calls.
Innovation Solution
An electronic device with a touch panel, camera, and processor performs face tracking by receiving a touch signal to enter face tracking mode, identifying regions of interest in captured frames, generating a target frame, and transmitting it to another device for display during video calls.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If face tracking function is added to video call application, then video quality and user interaction are improved, but device complexity and processing requirements increase
Solution Approach 1:
The patent divides the video frame processing into multiple stages: initial face detection to identify face regions, extraction of face images from detected regions, and separate processing of face images versus background. This segmentation allows the system to focus computational resources on face tracking while maintaining overall video quality, resolving the contradiction between improved video quality and reduced processing requirements.
2Measurement precision
If face tracking is performed on captured frames, then face detection accuracy is improved, but processing time and energy consumption increase
Solution Approach 1:
The patent performs preliminary face detection on the entire frame to identify potential face regions before extracting and processing individual face images. This preliminary action pre-identifies areas of interest, so that subsequent face tracking processing can be focused and optimized, reducing overall processing time while maintaining high detection accuracy.
Solution Approach 2:
The patent extracts face images from the captured frames based on detected face regions, separating face processing from general video processing. This extraction allows the system to apply specialized face tracking algorithms only to relevant portions of the frame, improving detection accuracy without requiring full-frame processing at maximum resolution.
3Adaptability or versatility
If multiple ROIs are combined to generate target frame, then face tracking functionality is enhanced, but device complexity increases
Solution Approach 1:
The patent combines multiple extracted face images (ROIs) into a single target frame that is then superimposed on the original video frame. This merging creates enhanced face tracking functionality where multiple faces can be tracked and displayed simultaneously, while the combination process is managed through standardized image processing operations that control complexity.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
A method for performing a face tracking function in an electric device is provided. The electric device has a touch panel, a camera, and a processor. The method includes the following steps. A touch signal is receiving by the touch panel. Under a video call, a face tracking mode is entered based on the touch signal by the processor. Face tracking is performed on a captured frame from the camera to obtain at least one region of interesting (ROI) of the captured frame by the processor, each of the ROI having an image of a face. A target frame is generated by combining the at least one ROI by the processor. The target frame is transmitted to another electric device by the processor, so that the target frame is shown on the another electric device as a video talk frame.