Texture Compression Using Two Hues and Modified Brightness
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing texture compression schemes for mobile devices, such as S3TC and iPACKMAN/ETC, face challenges in balancing hardware complexity and image quality, particularly in managing images with two distinct hues, where iPACKMAN/ETC introduces artifacts that degrade visual quality.
Innovation Solution
A method for encoding and decoding images by decomposing them into blocks with 2m×2n image elements, using color codewords and a color modifier to generate multiple color representations along specific directions in color space, allowing for efficient compression and decompression of images with two distinct hues.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If iPACKMAN/ETC compression scheme is used, then hardware complexity is reduced, but image quality deteriorates due to chrominance quantization artifacts
Solution Approach 1:
The image block is divided into two sub-blocks (first and second sub-blocks), each with its own base color. This segmentation allows differential encoding where only the difference between sub-blocks needs to be stored, reducing chrominance quantization artifacts while maintaining hardware efficiency.
Solution Approach 2:
The patent changes the encoding parameters by introducing differential color values (dcR, dcG, dB) that represent the difference between base colors of adjacent sub-blocks. This parameter transformation enables finer chrominance representation without increasing overall bit rate, thereby improving image quality while keeping hardware complexity manageable.
2Manufacturing precision
If S3TC compression scheme is used, then image quality is maintained, but hardware complexity and bandwidth usage increase
Solution Approach 1:
The patent extracts only the essential color information by representing each 2×2 sub-block with a single base color and storing only the differential values between adjacent sub-blocks. This extraction approach reduces the data structure complexity compared to S3TC's four-color palette while maintaining adequate image quality.
Solution Approach 2:
Instead of storing absolute color values for each pixel or complex palette information like S3TC, the patent inverts the approach by storing a base color and then storing only the differences (deltas) from that base. This inversion simplifies the hardware requirements while preserving color accuracy.
3Adaptability or versatility
If S3TC interpolation is performed, then color palette is generated, but multiplication operations increase hardware complexity
Solution Approach 1:
The patent uses simple addition operations instead of multiplication for color interpolation. The differential values are added to the base color to generate intermediate colors, which is a computationally cheaper operation that can be implemented with simpler hardware, reducing device complexity while maintaining the ability to generate color palettes.
Data Source
AI summary
In an image-encoding scheme, an input image is decomposed into image blocks comprising multiple image elements. The image blocks are then encoded into encoded blocks. An encoded block comprises a first color codeword, a second color codeword, a color modifier codeword and a color index sequence. The color codewords are representations of a first and second base color located on a first line in color space. The modifier codeword is a representation of at least one color modifier for modifying the first base color along a second line to obtain multiple color representations along the line. The second line has a different direction as compared to the first line. The index sequence comprises color indices associated with a color representation selected form i) the representations on the second line and ii) at least one representation based on the second base color.