Dual-Stage Image Compression for High Ratio and Quality
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image data processing methods in digital cameras face challenges in achieving high compression ratios without significantly deteriorating image quality, as higher compression ratios lead to increased distortion and loss of low-order bit information, particularly in areas with low pixel correlation like image edges.
Innovation Solution
The solution involves a dual-stage compression process where the first stage uses nonlinear transform compression and the second stage employs DPCM compression, both with distinct distortion characteristics, followed by reverse decompression processes to minimize visual deterioration and achieve a higher compression ratio.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If a single compression process (e.g., DPCM or nonlinear transform) is used to compress original image data, then the compression ratio can be increased, but image quality deteriorates due to compression distortion and loss of low-order bit information
Solution Approach 1:
The compression process is divided into multiple stages, each using a different compression method (DPCM for high-order bits, nonlinear transform for low-order bits). This segmentation allows each stage to optimize for its specific bit range, preventing the degradation that occurs when a single method is applied to all bits.
Solution Approach 2:
Different compression methods are applied to different parts of the image data based on their characteristics. High-order bits with high correlativity use DPCM, while low-order bits with low correlativity use nonlinear transform. This local quality approach ensures each bit range is compressed by the most suitable method, maintaining overall image quality while achieving high compression ratios.
2Quantity of substance
If the compression ratio is increased in existing methods, then more data can be stored in image memory, but compression distortion increases and image quality deteriorates
Solution Approach 1:
By segmenting the compression process into multiple stages with different methods, the system achieves higher overall compression ratios without the image quality deterioration that occurs with single-method compression. This enables more data to be stored in the same image memory capacity.
Solution Approach 2:
The patent combines multiple compression methods (DPCM and nonlinear transform) into a composite compression system. This composite approach leverages the strengths of each method while compensating for their weaknesses, achieving high compression ratios while maintaining image quality.
3Loss of information
If DPCM process is applied to high order bits with high correlativity, then high compression ratio is achieved, but the process is not effective for low order bits with low correlativity causing image distortion
Solution Approach 1:
The patent applies different compression methods to different bit ranges based on their correlation characteristics. High-order bits with high spatial correlativity are compressed using DPCM, while low-order bits with low correlativity (especially in edge regions) are compressed using nonlinear transform. This local quality differentiation prevents the image distortion that would occur if DPCM were applied uniformly to all bits.
Solution Approach 2:
The compression process is segmented into at least two stages: a first compression stage using DPCM for high-order bits, and a second compression stage using nonlinear transform for low-order bits. This segmentation allows each stage to be optimized for its specific domain, achieving high compression ratios while preserving image quality in challenging areas like edges.
Data Source
AI summary
Image data are compressed such that they do not visually deteriorate. Nonlinear compression and DPCM compression are performed in series. A compressing unit 41 that uses nonlinear transform inputs image data (L bits per pixel) of each color component. Image data are compressed to M bits (M<L) according to a compression transform tale TB1 that has been set up according to image information extracted from original image data. The compressing unit 41 performs compression transform having a characteristic similar to a gamma curve characteristic used in a gamma correction disposed downstream of the compressing unit 41. Image data compressed and transformed to M bits by the nonlinear compressing unit 41 are input to a DPCM compressing unit 42. The DPCM compressing unit 42 finally compresses input data to N bits (for example, N=10) according to a quantization table TB2. The compressed image data are stored to an image memory through a packing section 43.