Multi-Layer Image Compression with Floating-Point SNR Layer Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image compression technologies, such as JPEG-2000, face limitations in achieving high-quality wide-range multi-layer image compression due to issues with quantization, bit allocation, and spatial frequency separation, particularly in preserving the visual essence of images and handling wide dynamic ranges.
Innovation Solution
The use of 32-bit floating-point computations for pixel and transformed pixel values, adjustable floating-point deadbands, optimal band-split filter phases, and targeted SNR layers with adjustable quantization, allowing for efficient compression of wide-range images and improved spatial scalability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If JPEG-2000 DWT 9/7 bi-orthogonal subband configuration is used for compression, then compression ratio is improved, but image quality and visual essence preservation deteriorate
Solution Approach 1:
The patent changes the fundamental parameter of numerical representation from fixed-point integer arithmetic to floating-point arithmetic throughout the compression pipeline. This parameter change enables precise control of quantization steps and allows the system to maintain high image quality even at high compression ratios by dynamically adjusting quantization parameters without the precision losses inherent in integer-based systems.
Solution Approach 2:
The patent introduces dynamic adjustment of quantization parameters and deadband thresholds based on local image characteristics and desired quality levels. Unlike static quantization schemes, the system dynamically adapts precision requirements across different regions and compression levels, allowing optimal balance between compression ratio and visual quality preservation.
2Productivity
If quantization is applied to achieve compression, then compression efficiency is improved, but image fidelity and visual quality deteriorate
Solution Approach 1:
The patent applies different quantization precision levels to different regions of the image based on local characteristics such as edge density, texture complexity, and visual importance. Critical regions receive finer quantization to preserve fidelity, while less critical regions use coarser quantization to maximize compression efficiency, achieving an optimal local balance between these competing goals.
Solution Approach 2:
The patent introduces a deadband mechanism that applies quantization selectively - values within the deadband threshold are set to zero without further quantization, while values outside the deadband receive full quantization processing. This partial application of quantization reduces information loss for small coefficient values while maintaining compression efficiency for large coefficients.
3Speed
If fixed-point arithmetic is used for compression computations, then computational speed is improved, but precision and handling of wide dynamic range deteriorate
Solution Approach 1:
The patent substitutes the mechanical fixed-point arithmetic system with a floating-point arithmetic system throughout the compression pipeline. This substitution enables precise handling of wide dynamic ranges in image data, accurate representation of small and large coefficient values, and flexible control of quantization parameters, overcoming the precision limitations of fixed-point systems while modern hardware makes the speed penalty acceptable.
4Productivity
If DWT transform is used for frequency separation, then compression capability is improved, but handling of wide dynamic range and precision deteriorate
Solution Approach 1:
The patent changes the numerical representation parameter from fixed-point to floating-point throughout the DWT transform and subsequent processing stages. This enables the transform to accurately handle the full dynamic range of image data without precision clipping, maintains accurate representation of small high-frequency coefficients, and allows precise control of quantization parameters at each subband level.
Data Source
AI summary
Systems, methods, and computer programs for high quality wide-range multi-layer image compression coding, including consistent ubiquitous use of floating point values in essentially all computations; an adjustable floating-point deadband; use of an optimal band-split filter; use of entire SNR layers at lower resolution levels; targeting of specific SNR layers to specific quality improvements; concentration of coding bits in regions of interest in targeted band-split and SNR layers; use of statically-assigned targets for high-pass and/or for SNR layers; improved SNR by using a lower quantization value for regions of an image showing a higher compression coding error; application of non-linear functions of color when computing difference values when creating an SNR layer; use of finer overall quantization at lower resolution levels with regional quantization scaling; removal of source image noise before motion-compensated compression or film steadying; use of one or more full-range low bands; use of alternate quantization control images for SNR bands and other high resolution enhancing bands; application of lossless variable-length coding using adaptive regions; use of a folder and file structure for layers of bits; and a method of inserting new intra frames by counting the number of bits needed for a motion compensated frame.