Video Call Face Recognition via Shared Media Channel
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing face recognition methods in voice/video calls require installing a dedicated face recognition application on terminals, consuming additional bandwidth and port resources, and often necessitate complex operations and interruptions.
Innovation Solution
Utilize the existing video call media transmission channel to transmit face video streams, eliminating the need for a separate transmission channel and application installation by integrating face recognition into the call process through a media server.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a user equipment simultaneously performs channel estimation and signal decoding, then communication efficiency is improved, but processing complexity and power consumption increase
Solution Approach 1:
The patent divides the signal processing tasks into separate threads: a first thread dedicated to channel estimation and a second thread dedicated to signal decoding. This segmentation allows each thread to specialize in one function, improving overall efficiency while managing complexity through modular architecture.
Solution Approach 2:
The patent performs channel estimation before signal decoding by having the first thread complete its estimation task and pass results to the second thread. This preliminary action ensures that decoding can proceed with accurate channel information, improving overall communication efficiency.
2Productivity
If a user equipment simultaneously performs channel estimation and signal decoding, then communication efficiency is improved, but power consumption increases
Solution Approach 1:
The patent segments the processing workload into separate threads that can execute independently and efficiently. The first thread handles channel estimation while the second thread handles signal decoding, allowing the system to optimize power consumption for each specific task rather than inefficiently processing both simultaneously with a single thread.
Solution Approach 2:
By completing channel estimation as a preliminary action before signal decoding, the system avoids redundant processing and ensures that decoding operations can proceed with pre-computed channel information, reducing overall power consumption compared to simultaneous processing.
3Measurement precision
If channel estimation is performed before signal decoding, then decoding accuracy is improved, but processing time increases
Solution Approach 1:
The patent implements channel estimation as a preliminary action that completes before signal decoding begins. The first thread finishes estimation and passes results to the second thread, ensuring decoding accuracy while the sequential structure manages processing time through efficient task coordination.
Solution Approach 2:
By segmenting the processing into separate threads with the first thread dedicated to estimation and the second to decoding, the system optimizes both accuracy and time management. The segmentation allows each thread to execute its specific task efficiently without resource contention.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A communication method, apparatus, and system are provided, which relate to the field of communication technologies, to reduce bandwidth resources and port resources of a terminal that are occupied by performing face recognition in a call process. The method includes: establishing a video call media transmission channel used to transmit a call video stream between a call terminal and a peer call terminal in a video call service, where the call video stream includes video content photographed by the call terminal or the peer call terminal; receiving, from a media server, a SIP message including a face recognition request identifier, where the face recognition request identifier is for requesting to perform face recognition on a user corresponding to the call terminal; sending a response message of the SIP message to the media server, where the response message of the SIP message indicates that the user corresponding to the call terminal agrees to perform face recognition; next sending a face video stream to the media server through the video call media transmission channel, where the face video stream includes a face image of the user corresponding to the call terminal; and further receiving a face recognition result from the media server.