Layer-based video encoding with polynomial transforms for text legibility
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional video codecs face challenges in balancing legibility and compression ratio, particularly when encoding video signals with detailed content like text, as aggressive quantization can compromise legibility while less aggressive quantization reduces compression efficiency.
Innovation Solution
A layer-based encoding and decoding technique that lossily compresses video signals into multiple residual layers, using masks to distinguish relevant 'care' pixels from irrelevant 'don't-care' pixels, and selectively applies a polynomial transform instead of traditional frequency transforms for efficient compression.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the encoder quantizes the DCT coefficients aggressively to reduce the size and number of non-zero DCT coefficients, then compression ratio is improved, but legibility of reconstructed text suffers
Solution Approach 1:
The patent segments the video content into multiple layers based on motion characteristics. The foreground layer containing text and other high-contrast features is separated from the background layer. This segmentation allows different compression strategies to be applied to each layer, preserving text legibility in the foreground layer while achieving high compression ratios for the background layer.
Solution Approach 2:
The patent applies different quality levels to different regions of the video content. The foreground layer, which contains text and high-contrast features, is encoded with higher quality and less aggressive quantization to maintain legibility. The background layer is encoded with lower quality and more aggressive quantization to maximize compression ratio. This local quality differentiation resolves the contradiction between overall compression efficiency and text legibility.
2Manufacturing precision
If the encoder quantizes the DCT coefficients less aggressively to maintain legibility, then legibility of reconstructed text is improved, but compression ratio suffers
Solution Approach 1:
The patent segments the video content into multiple layers based on motion characteristics. The foreground layer containing text and other high-contrast features is separated from the background layer. This segmentation allows different compression strategies to be applied to each layer, preserving text legibility in the foreground layer while achieving high compression ratios for the background layer.
Solution Approach 2:
The patent applies different quality levels to different regions of the video content. The foreground layer, which contains text and high-contrast features, is encoded with higher quality and less aggressive quantization to maintain legibility. The background layer is encoded with lower quality and more aggressive quantization to maximize compression ratio. This local quality differentiation resolves the contradiction between overall compression efficiency and text legibility.
3Ease of manufacture
If conventional codecs process all pixels in a block equally, then encoding simplicity is maintained, but computational burden increases and compression efficiency decreases for partially masked content
Solution Approach 1:
The patent segments the pixel array into care pixels and don't-care pixels using a mask. This segmentation allows the encoder to focus computational resources only on the care pixels that contribute to the reconstructed output, while ignoring the don't-care pixels. This significantly reduces the computational burden and improves compression efficiency for partially masked content while maintaining encoding simplicity through the use of standard transform operations on the reduced pixel set.
Data Source
AI summary
A technique for processing video includes receiving a pixel array, such as a block or layer of video content, as well as a mask that distinguishes masked, “don't-care” pixels in the pixel array from unmasked, “care” pixels. The technique encodes the pixel array by taking into consideration the care pixels only, without regard for the don't-care pixels. An encoder operating in this manner can produce a simplified encoding of the pixel array, which represents the care pixels to any desired level of precision, without regard for errors in the don't-care pixels, which are irrelevant to reconstruction. Further embodiments apply a polynomial transform in place of a frequency transform for encoding partially-masked video content, and/or video content meeting other suitable criteria.