Adaptive Chroma Subsampling for Video Bandwidth Optimization

The patent applies different chroma subsampling resolutions to different regions of the image based on their uniformity characteristics. Low-uniformity regions (containing edges, textures, or important details) receive high-resolution 4:4:4 subsampling to preserve quality, while high-uniformity regions (solid colors, gradients) use low-resolution 4:2:0 subsampling to reduce bandwidth. This local adaptation resolves the contradiction by maintaining image quality only where necessary.

The image is divided into multiple tiles, and each tile is further segmented into uniformity regions that are classified and processed with appropriate subsampling resolutions. This segmentation allows the system to apply quality-preserving high-resolution subsampling only to specific segments (low-uniformity regions) while using compression-friendly low-resolution subsampling in other segments, thereby reducing overall bandwidth usage without sacrificing critical image quality.

By classifying regions according to their uniformity and applying appropriate subsampling resolutions locally, the patent reduces the total computational workload during encoding and decoding. Low-uniformity regions requiring quality preservation use 4:4:4 subsampling, while high-uniformity regions use the faster-to-process 4:2:0 subsampling, thereby reducing overall processing time while maintaining quality where needed.

Instead of applying full high-resolution subsampling uniformly across the entire image (excessive action), the patent applies high-resolution processing only partially to regions where it is actually necessary (low-uniformity regions). This partial action approach reduces encoding and decoding time by avoiding unnecessary high-resolution processing in regions where it would not improve perceived quality.

The patent strategically applies low-resolution chroma subsampling (4:2:0) only to high-uniformity regions where quality loss is imperceptible, while maintaining high-resolution 4:4:4 subsampling in low-uniformity regions where quality matters. This local differentiation allows the system to achieve high network transmission efficiency by reducing bandwidth in appropriate regions without deteriorating overall image quality.

The patent dynamically changes the chroma subsampling parameter (resolution) based on the local uniformity characteristics of different image regions. By adapting the subsampling parameter to match regional characteristics, the system achieves efficient bandwidth utilization in high-uniformity areas while preserving quality in low-uniformity areas, thus improving network transmission efficiency without sacrificing image quality.

The patent implements a relatively simple uniformity classification mechanism that divides the image into regions and assigns appropriate subsampling resolutions based on basic uniformity metrics. This local classification approach achieves bandwidth efficiency without requiring complex adaptive algorithms, as the classification rules are straightforward and computationally lightweight.

The patent changes the chroma subsampling parameter based on pre-computed uniformity metrics rather than through complex real-time adaptation. The uniformity classification provides a simple framework for determining appropriate subsampling levels, avoiding the need for sophisticated device complexity while still achieving bandwidth efficiency through parameter variation.