Motion-Based Adaptive Quantization for Video Streaming
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image compression technologies face challenges in reconciling the opposing requirements of maintaining high quality for static content and reducing bandwidth for dynamic content in video streaming, particularly in 'desktop streaming' scenarios where still portions require sharper contrasts and moving content needs less visual quality.
Innovation Solution
The implementation of motion-based adaptive quantization, where the quantization parameter (QP) for image data is adjusted from a start QP to a target QP in a multi-step change, specifically increasing compression quality for motionless portions and decreasing it for non-motionless portions, allowing for customized compression based on intra or inter prediction modes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If a single quantization parameter is used for all regions, then device complexity is reduced, but image quality for static regions deteriorates
Solution Approach 1:
The patent applies different quantization parameters to different regions of the image based on motion characteristics. Static regions (where motion vectors are zero or similar to neighboring blocks) use a first QP value optimized for quality, while non-static regions use a second QP value optimized for compression efficiency. This local differentiation resolves the contradiction by providing high quality where needed without unnecessarily complicating the overall system.
Solution Approach 2:
The patent segments the image into static and non-static regions based on motion vector analysis. By dividing the image content into these two categories and applying different compression parameters to each segment, the system achieves both simplicity (through automated segmentation based on clear criteria) and high image quality (through targeted parameter application to static regions).
2Quantity of substance
If quantization parameter is increased to reduce bit rate, then bandwidth consumption is reduced, but image quality deteriorates
Solution Approach 1:
The patent applies a higher QP value (lower quality, higher compression) only to non-static regions where motion is present and visual quality is less critical. Static regions receive a lower QP value (higher quality, lower compression) to preserve sharpness and detail. This selective approach reduces overall bit rate while maintaining acceptable image quality by concentrating compression efforts where they are least noticeable.
3Manufacturing precision
If quantization parameter is decreased to improve image quality, then image quality improves, but bandwidth consumption increases
Solution Approach 1:
The patent decreases the QP value (improving quality) specifically for static regions where visual fidelity is most important, while maintaining a higher QP value (lower quality, lower bandwidth) for non-static regions. This localized quality enhancement achieves the goal of improved image quality without proportionally increasing bandwidth consumption, as the high-quality encoding is applied only where it provides the most benefit.
Data Source
AI summary
A method and apparatus for compressing a data stream comprising a plurality of pictures are described. A first quantization parameter (QP) from a plurality of QPs is determined, for a static region in a current picture. The plurality of QPs change in accordance with a multi-step change from a start QP to a target QP and each one of the plurality of QPs is to be applied to a respective one from successive static regions in successive pictures. In response to determining that the static region is to be compressed based upon an inter prediction mode, compression of the static region is caused based upon the first QP; and in response to determining that the static region is to be compressed based upon an intra prediction mode, compression of the static region is caused based upon a minimum QP.