Image Encoding Apparatus Using Adaptive DPCM and DCT Modes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image compression methods face inefficiencies, particularly in encoding and decoding images with texture areas lacking similarity and regularity, leading to increased image data size and loss during quantization processes.
Innovation Solution
A method and apparatus for selectively entropy encoding residual values of image access units, determining if they are the same as previously encoded units, and encoding only the number of repetitions or residual values accordingly, using techniques like exponential-Golomb codes and Huffman coding to optimize compression.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If lossless image encoding is used without discrete cosine conversion and quantization, then image quality is preserved without loss, but the image data size increases when encoding texture areas with no similarity and regularity
Solution Approach 1:
The patent changes the encoding parameter selection based on the characteristics of the image region. For textureless areas with high similarity, it uses differential pulse code modulation (DPCM) with run-length encoding. For texture areas with low similarity, it switches to discrete cosine transformation (DCT) with quantization. This dynamic parameter change resolves the contradiction by adapting the encoding method to the local image characteristics.
Solution Approach 2:
The encoding system dynamically switches between different encoding modes (lossless DPCM and lossy DCT) based on the measured similarity and regularity of pixel values in each block. This dynamic adaptation allows the system to optimize compression ratio for each region while maintaining acceptable image quality, resolving the contradiction between image quality preservation and data size reduction.
2Quantity of substance
If discrete cosine conversion and quantization are used to improve compression rate, then image data size is reduced, but loss is generated in the image
Solution Approach 1:
The patent applies different encoding quality levels to different regions of the image. Textureless areas are encoded with lossless DPCM to preserve quality, while texture areas are encoded with DCT and quantization to achieve better compression. This local differentiation resolves the contradiction by applying lossy compression only where it is less perceptible.
Solution Approach 2:
The system dynamically selects between lossless and lossy encoding methods based on the local image characteristics. By measuring the similarity and regularity of pixel values, the system determines the appropriate encoding mode for each block, allowing it to minimize overall image data size while maintaining acceptable quality in critical regions.
3Productivity
If image is encoded in units of blocks using quantization, then compression is achieved, but loss is generated during the quantization process
Solution Approach 1:
The patent changes the encoding parameters (encoding method and quantization level) based on the characteristics of each image block. Blocks with high pixel similarity use DPCM with minimal or no quantization, while blocks with low similarity use DCT with appropriate quantization levels. This parameter adaptation resolves the contradiction by optimizing compression rate for each block while minimizing quality loss.
Solution Approach 2:
The image is divided into blocks that are further segmented into textureless areas and texture areas based on pixel similarity measurements. This segmentation allows different encoding strategies to be applied to different segments, achieving overall compression while preserving quality in segments where it matters most.
4Reliability
If lossless JPEG or JPEG-LS encoding is used to encode pixel values directly, then no quantization loss occurs, but image data size becomes larger for texture areas
Solution Approach 1:
The patent changes the encoding method parameter from pure lossless encoding to a hybrid approach that selects between lossless DPCM and lossy DCT based on image characteristics. This parameter change allows the system to achieve better compression rates in texture areas where lossless encoding is inefficient, while maintaining lossless quality in textureless areas.
Solution Approach 2:
The system dynamically adjusts the encoding approach based on the measured properties of each image block. By calculating pixel similarity and regularity, the system determines whether to use lossless or lossy encoding, dynamically optimizing the compression rate while maintaining acceptable quality standards for each region.
Data Source
AI summary
Provided is a method and apparatus for compression encoding/decoding an image based on the repetition of an access unit and the repetition of a residual value.