Video Frame User Presence Detection for Conference Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing multimedia multi-user collaboration applications, such as videoconferencing systems, face challenges in optimizing video performance due to the computational intensity of processing video frames for all participants, leading to resource burden and inefficiency, especially in large-scale meetings where only a few participants are actively engaged.
Innovation Solution
The system employs a video frame user presence detection system using a time-of-flight sensor to create a human object presence heat map, allowing for differential encoding of video frames based on user presence, with high-quality encoding for active users and lower computational burden for background portions, and sends this data out-of-band to reduce processing resources and data loss.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If video frames are processed with high-quality encoding for all participants, then video quality is improved, but resource consumption increases
Solution Approach 1:
The patent applies local quality by differentiating video processing quality based on user activity status. Active users receive high-quality encoding with full video frame processing, while inactive users receive lower-quality encoding with reduced processing. This resolves the contradiction by applying high resource consumption only where necessary (active users) rather than uniformly across all participants.
Solution Approach 2:
The patent segments users into two categories: active users and inactive users. This segmentation allows the system to apply different processing strategies to different groups, thereby reducing overall resource consumption while maintaining high video quality for those who need it most.
2Reliability
If video frames are processed for all participants, then completeness of video data is improved, but processing load increases
Solution Approach 1:
The patent extracts and processes only the necessary portions of video data based on user activity. For inactive users, the system extracts minimal essential information rather than processing complete video frames. This reduces processing load while maintaining sufficient video data completeness for effective communication.
3Manufacturing precision
If high-quality encoding is applied to all video frames, then video quality is improved, but data transmission volume increases
Solution Approach 1:
The patent applies local quality by transmitting high-quality video data only for active users while reducing data transmission quality for inactive users. This approach maintains video quality where needed while significantly reducing overall data transmission volume across the network.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach reduces resource consumption and improves video quality by prioritizing active user frames, enhancing the efficiency and performance of video conferencing sessions while minimizing processing load and data loss.
Implementation Method 1
a time-of-flight sensor or other distance sensor to provide distance data descriptive of the distance between a user's body and the video camera
Data Source
AI summary
An information handling system executing a multimedia multi-user collaboration application (MMCA) may include a memory; a power management unit; a video camera to capture video of a user participating in a video conference session conducted by a processor executing code instructions of the MMCA via a first network link on a network interface device; a time-of-flight (TOF) sensor to provide distance data descriptive of the distance between a user's body and the video camera; the processor configured to execute a video frame user presence detection system to: slice a video frame into a plurality of video frame slices; based on the distance data, create a human object presence (HOP) heat map to assign a probability of the presence of the user's body within each video frame slice of the video frame; the processor configured to select, among the plurality of video frame slices, those video frame slices with an assigned probability of presence of the user's body that exceeds a threshold probability as a HOP video frame slice group where the user's body appears within the video frame; and the network interface device configured to send data defining the HOP video frame slice group to a sink information handling system participating in the video conference session over the network to assist decoding of the video frame at the sink information handling system.