Adaptive Blur Detection for Video Streaming Bandwidth
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video conferencing technologies face bandwidth limitations due to high demands for streaming video content, especially in remote working environments with inconsistent connections, and existing compression solutions offer only marginal bandwidth savings at the expense of video quality.
Innovation Solution
An AI/ML-based system that assembles and transmits a set of reference frames covering different facial expressions or object attributes, using a Multi-Layer Perceptron encoder to decide which frames to cache, and employs an adaptive threshold for blur detection to optimize bandwidth usage by distinguishing between blurred and non-blurred frames.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If video compression solutions are employed to reduce bandwidth requirements, then bandwidth consumption is reduced, but video quality deteriorates
Solution Approach 1:
The system performs preliminary action by pre-assembling and caching reference frames that cover different facial expressions and object attributes before transmission. This allows the receiver to reconstruct video frames more accurately without requiring high bandwidth for transmitting every frame detail, thus reducing bandwidth consumption while maintaining video quality.
Solution Approach 2:
The system creates copies of reference frames with different attributes (facial expressions, object attributes) and transmits only these essential reference copies rather than all video frames. The receiver uses these copied reference frames to reconstruct the actual video content, significantly reducing bandwidth requirements while preserving visual quality.
2Manufacturing precision
If all video frames are transmitted to maintain video quality, then video quality is maintained, but bandwidth requirements increase
Solution Approach 1:
The system extracts and transmits only the essential reference frames that contain key information about facial expressions and object attributes, separating these from the full video stream. By taking out only the necessary reference information rather than transmitting all frames, the system maintains video quality while reducing bandwidth requirements.
Solution Approach 2:
Reference frames are pre-assembled and cached in advance based on predicted video content needs. This preliminary action allows the system to have essential reference information ready before it is needed for reconstruction, eliminating the need to transmit redundant frame data during real-time streaming.
3Loss of energy
If reference frames are cached to reduce transmission data, then bandwidth usage is reduced, but video reconstruction accuracy may deteriorate
Solution Approach 1:
The system applies local quality by creating reference frames with different specific attributes (facial expressions, object attributes) tailored to local content requirements. Each cached reference frame is optimized for specific local conditions, ensuring that when a particular reference frame is used for reconstruction, it provides high accuracy for that specific content type.
Solution Approach 2:
The system changes parameters by varying the attributes of cached reference frames (different facial expressions, object attributes, time periods). This allows the receiver to select the most appropriate reference frame based on current video content parameters, maintaining high reconstruction accuracy while using limited bandwidth.
Data Source
AI summary
Systems and methods herein address reference frame selection in video streaming applications using one or more processing units to identify a frame of a sequence of frames as a blurred frame based at least in part on a first variance of motion (VoM) of the frame being less than or equal to an adaptive threshold that is based in part on a moving average of variance of motion (MAoV) determined using one or more reference frames.