Color Image Compression via Component Line Packetization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for compressing and decompressing color images are inefficient in managing increasing video bandwidth demands due to high video resolutions, frame rates, and stream numbers, particularly in the context of the JPEG XS standard, which requires effective packetization and transformation of image samples.
Innovation Solution
A method involving a decorrelative transform followed by Digital Wavelet Transform (DWT) decompositions and entropy coding, where samples are grouped into packets based on a sampling pattern, with specific filter-bank operations and reference maintenance to optimize compression and decompression efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If current JPEG XS standard packetization methods are used, then image compression is achieved, but buffer sizes and metadata overhead become excessively large
Solution Approach 1:
The patent segments the packed data set into smaller packets organized by component and line, rather than using large buffers. Each packet contains a specific portion of the transformed data, allowing for reduced buffer sizes while maintaining compression efficiency through structured organization of the segmented data
Solution Approach 2:
The patent introduces a new dimension in data organization by creating a hierarchical packet structure based on component indices and line indices. This multi-dimensional organization (component × line × sample) replaces the traditional single-dimension buffer approach, enabling more efficient memory utilization and reduced overhead
2Measurement precision
If more packets are created for detailed image representation, then image quality is improved, but metadata overhead increases
Solution Approach 1:
The patent extracts and eliminates redundant metadata from the packet structure. By organizing packets efficiently and removing unnecessary header information, the method reduces metadata overhead while preserving all essential information needed for image reconstruction, thus maintaining image quality without the penalty of excessive metadata
3Manufacturing precision
If extensive reorder buffers are used to maintain sample positions, then spatial accuracy is improved, but hardware complexity increases
Solution Approach 1:
The patent performs preliminary organization of samples during the forward transform process, assigning each sample its final position metadata before packetization. This preliminary action eliminates the need for extensive reorder buffers at the decoding stage, as samples are already positioned correctly in the packet structure, thereby reducing hardware complexity while maintaining spatial accuracy
Data Source
AI summary
The disclosure relates to a method for compressing in a packed data set a color image comprising color image samples arranged in a reference grid according to a sampling pattern, and comprising the steps of (a) performing a decorrelative transform providing component images, (b) performing a number of discrete wavelet decompositions on said component images and (c) forming packets in the packed data set by processing each consecutive line in the reference grid and each component and grouping all samples of said component in said line in one or more packets, and appending said one or more packets to the packed data set. The disclosure also relates to a device for compressing, a computer program for compressing, a method for decompressing, a device for decompressing, and a computer program for decompressing. The disclosure also relates to the packed data set produced by said method for compressing.