《数字图像处理》课程教学课件（Digital Image Processing）图像压缩 ——霍夫曼-算术-LZW编码

Digital Image Processing Chapter 8: Image Compression-2

Digital Image Processing Chapter 8: Image Compression-2

Outline 1 Fundamentals 2 Some basic compression method ·Huffman coding ·Arithmetic coding ·LZW coding Block Transform coding ·Predictive coding

Outline 1 Fundamentals 2 Some basic compression method • Huffman coding • Arithmetic coding • LZW coding • Block Transform coding • Predictive coding • …

Error-Free Compression:Huffman Coding Huffman coding:give the smallest possible number of code symbols per source symbols. Step 1:Source reduction Original source Source reduction Symbol Probability 1 2 3 4 a 0.4 0.4 0.4 0.40.6 as 0.3 0.3 0.3 0.3 0.4 a 0.1 0.10.2 T0.3 4 0.1 0.1 0.1 as 0.06 0.1 as 0.04

Error-Free Compression: Huffman Coding Step 1: Source reduction Huffman coding: give the smallest possible number of code symbols per source symbols

Error-Free Compression:Huffman Coding Step 2:Code assignment procedure Original source Source reduction Sym. Prob. Code 1 2 3 4 2 0.4 1 0.41 0.41 0.41 -0.60 as 0.3 00 0.300 0.300 0.300 0.41 a 0.1 011 0.1011 -0.2010-0.301← a4 0.1 0100 0.10100 0.1011 a3 0.06 010104 -0.101014 as 0.04 010114 The code is instantaneous uniquely decodable without referencing succeeding symbols

Error-Free Compression: Huffman Coding Step 2: Code assignment procedure The code is instantaneous uniquely decodable without referencing succeeding symbols

Arithmetic Coding Nonblock code:one-to-one correspondence between source symbols And code words does not exist. Concept:The entire sequences of source symbols is assigned a single arithmetic code word in the form of a number in an interval of real number between 0 and 1. Encoding sequence 41 9 0.2 0.08 0.072 0.0688 a4 0.06752 3 a 3 02 h 2 子 a a a 0.04 0.056 0.0624

Arithmetic Coding Nonblock code: one-to-one correspondence between source symbols And code words does not exist. Concept: The entire sequences of source symbols is assigned a single arithmetic code word in the form of a number in an interval of real number between 0 and 1

Arithmetic Coding Example 0.2x0.40.04+0.8x0.04 0.056+0.8x0.016 Encoding sequence 01 03 The number 0.2 0.08 0.072 0.0688 as between 0.0688 0.06752 and0.06752 0 0 d3 can be used to represent the a 3 sequence 0 0.04 0.056 0.0624 al a2 a,a3 a4 0.2x0.2 0.04+0.4x0.04 0.056+0.4x0.016 Source Symbol Probability Initial Subinterval 41 0.2 [0.0,0.2) az 0.2 0.2,0.4) a 0.4 0.4,0.8) aa 0.2 [0.8.1.0)

Arithmetic Coding Example 0.2x0.4 0.04+0.8x0.04 0.056+0.8x0.016 0.2x0.2 0.04+0.4x0.04 0.056+0.4x0.016 The number between 0.0688 and 0.06752 can be used to represent the sequence a1 a2 a3 a3 a4

LZW Coding Lempel-Ziv-Welch coding:assign fixed length code words to variable length sequences of source symbols. Dictionary Location Entry 0 0 1 255 255 256 511 -

LZW Coding Lempel-Ziv-Welch coding : assign fixed length code words to variable length sequences of source symbols

LZW Coding Currently Dictionary Recognized Pixel Being Encoded Location Sequence Processed Output (Code Word) Dictionary Entry 39 39 39 39 256 39-39 39 126 39 257 39-126 126 126 126 258 126-126 126 39 126 259 126-39 39 39 39-39 126 256 260 39-39-126 126 126 126-126 9 258 261 126-126-39 39 39 30-39 126 39-39-126⊙ 4---126 260 262 39-39-126-126 24 Bits 126 39 126-39 39 259 263 126-39-39 30 126 39-126 126 257 264 39-126-126 9 Bits 126 126

LZW Coding 24 Bits 9 Bits