In this thesis various methods for lossless compression of source image data are analyzed and discussed. The main focus in this work is lossless compression algorithms based on context modeling using tree structure.
We are going to compare CALIC, GCT-I algorithms to the
JPEG2000 standard algorithm which will be considered the reference of comparison.
This work includes research on how to modify CALIC algorithm in continuous-tone mode by truncating tails of the error histogram which may lead to improve CALIC compression performance.
Also, we are going to propose a modification to CALIC in binary
mode by eliminating error feedback mechanism. As when any pixel to be encoded has a different grey level than any of the neighboring pixels, CALIC triggers an escape sequence that switches the algorithm from binary mode to continuous-tone mode. Which means in this case the pixel will be treated as if it was in continuous-tone region. This minor modification should improve CALIC performance in binary images.
Finally, we are going to discuss the GCT-I on medical images and compare results to the JPEG2000 standard.
Table of Contents
1. INTRODUCTION
1.1 Data compression
1.2 Basic concepts in image compression
1.2.1 Lossless and lossy cases
1.2.2 Measures of compression
1.2.3 Paradigm of compression
1.2.4 Arithmetic coding
1.3 Thesis objectives
1.4 Contents of thesis' chapters
2. CONTEXT MODELING
2.1 Fundamentals
2.2 Context Tree
2.2.1 Structure of context tree
2.2.2 Static and semi-adaptive approaches
2.2.3 Construction of an initial context tree
2.2.4 Pruning of context tree
3. LOSSLESS IMAGE COMPRESSION TECHNIQUES
3.1 JPEG2000 standard
3.2 CALIC
3.2.1 CALIC Frame Structure
3.2.2 Neighborhood pixels involved
3.2.3 LINEAR PREDICTION - GAP
3.2.4 CODING CONTEXT
3.2.5 CONTEXT MODELING FOR ADAPTIVE ERROR FEEDBACK
i. Contexts Formation
ii. Error Feedback
3.2.6 ENTROPY CODING OF PREDICTION ERRORS
3.2.6.1 Error Sign Flipping
3.2.6.2 Remapping Errors
3.2.6.3 Histogram Tail Truncation
A. Proposed modification
3.2.7 Continuous-tone mode results
3.2.8 BINARY MODE
A. Proposed modification
3.3 General Context Tree based on Intensity
3.3.1 Context template
3.3.2 Compression scheme
3.3.3 GCT-I Simulation results
3.4 COMPARATIVE ANALYSIS OF CONSIDERED TECHNIQUES
4. CONCLUSIONS AND SUGGESTION FOR FUTURE WORK
4.1 Conclusions
4.2 Suggestion for Future Work
Objectives & Research Topics
The primary objective of this thesis is to analyze and compare various lossless compression algorithms, with a specific focus on those utilizing context modeling through tree structures. The research explores methods to improve performance in both continuous-tone and binary modes by proposing modifications to existing algorithms, such as CALIC, and evaluating the GCT-I algorithm on medical imaging datasets.
- Comparison of standard lossless image compression techniques (JPEG2000, CALIC, GCT-I).
- Development of modifications to the CALIC algorithm for continuous-tone mode via error histogram truncation.
- Refinement of CALIC in binary mode by eliminating the error feedback mechanism.
- Performance evaluation and comparative analysis on natural smooth and medical images.
Excerpt from the Book
1.1 Data compression
In computer science and information theory, data compression or source coding is the process of encoding information using fewer bits than an un-encoded representation would use through use of specific encoding schemes. For example, any article could be encoded with fewer bits if one were to accept the convention that the word "compression" be encoded as "comp". One popular instance of compression that many computer users are familiar with is the ZIP file format, which, as well as providing compression, acts as an archiver, storing many files in a single output file.
As is the case with any form of communication, compressed data communication only works when both the sender and receiver of the information understand the encoding scheme. For example, this text makes sense only if the receiver understands that it is intended to be interpreted as characters representing the English language. Similarly, compressed data can only be understood if the decoding method is known by the receiver. Some compression algorithms exploit this property in order to encrypt data during the compression process so that decompression can only be achieved by an authorized party (e.g. through the use of a password).
Compression is useful because it helps reduce the consumption of expensive resources, such as disk space or transmission bandwidth. On the downside, compressed data must be uncompressed to be viewed (or heard), and this extra processing may be detrimental to some applications. For instance, a compression scheme for video may require expensive hardware for the video to be decompressed fast enough to be viewed as it's being decompressed (you always have the option of decompressing the video in full before you watch it, but this is inconvenient and requires storage space to put the uncompressed video). The design of data compression schemes therefore involve trade-offs between various factors, including the degree of compression, the amount of distortion introduced (if using a lossy compression scheme), and the computational resources required to compress and uncompress the data.
Summary of Chapters
1. INTRODUCTION: This chapter introduces data compression concepts, distinguishes between lossless and lossy methods, and outlines the thesis objectives and scope.
2. CONTEXT MODELING: This section details the theoretical fundamentals of context modeling, including the structure, construction, and pruning of context trees for probability estimation.
3. LOSSLESS IMAGE COMPRESSION TECHNIQUES: This chapter provides a comprehensive analysis of JPEG2000, CALIC, and GCT-I, proposing specific modifications to improve their performance on different image types.
4. CONCLUSIONS AND SUGGESTION FOR FUTURE WORK: This final chapter summarizes the research findings, highlights the performance improvements achieved by the proposed modifications, and offers suggestions for future developments.
Keywords
Lossless compression, Context modeling, Image compression, CALIC, GCT-I, JPEG2000, Error feedback, Entropy coding, Histogram truncation, Data compression, Image processing, Medical images, Predictive coding, Bit-rate, Context tree.
Frequently Asked Questions
What is the core focus of this thesis?
The thesis focuses on analyzing and comparing lossless image compression algorithms that rely on context modeling and tree-based structures.
Which compression algorithms are examined?
The research primarily evaluates and compares the JPEG2000 standard, the CALIC (Context-based, Adaptive, Lossless Image Codec), and the GCT-I (General Context Tree based on Intensity) method.
What is the primary goal of the study?
The main goal is to improve the compression efficiency of existing methods, specifically by modifying the CALIC algorithm for continuous-tone and binary image modes and assessing GCT-I on medical datasets.
What methodology is employed for these comparisons?
The study uses experimental simulation on specific test sets (natural smooth images and medical images) and evaluates performance based on criteria such as bit-rate (bits per pixel) and compression efficiency.
What does the main part of the work cover?
The main part covers the theory of context modeling, a detailed examination of JPEG2000 and CALIC architectures, the implementation of histogram tail truncation and feedback elimination, and the development of the GCT-I approach.
Which keywords define this work?
The work is characterized by terms such as Lossless compression, Context modeling, CALIC, GCT-I, Error feedback, and Entropy coding.
Why is the error feedback mechanism in CALIC problematic for binary images?
The error feedback mechanism is designed for continuous-tone images; when triggered in binary-like regions (where pixels have widely separated gray levels), it can lead to a decrease in the compression ratio.
How does the proposed modification for binary mode work?
The study proposes eliminating the positive feedback mechanism specifically when the algorithm triggers an escape sequence due to a different gray level in binary images, thereby improving overall compression performance.
What is the specific advantage of the GCT-I algorithm?
GCT-I is particularly effective for medical images, which are characterized by a small number of colors and sharp color transitions where traditional linear predictive techniques often fail.
- Citation du texte
- Assistant Teacher Mohammad El-Ghoboushi (Auteur), 2014, Comparison of Lossless Image Compression Techniques based on Context Modeling, Munich, GRIN Verlag, https://www.grin.com/document/294973
