From Webresearch

Contents 1 Introduction and Definitions 1.1 Lossless 1.2 Lossy 1.3 Universal Source Coding 1.4 Network Source Coding 2 Algorithms 2.1 Lossless 2.2 Lossy 2.3 Universal 3 Source Coding Limits, Bounds, and Inequalities

Introduction and Definitions

Lossless

Lossy

Universal Source Coding

Network Source Coding

The field of network source coding is the portion of multiterminal information theory that generalizes source coding from the point-to-point network studied by Shannon to networks with multiple transmitters, multiple receivers, or side information available at one or more nodes in the network. Examples of lossless network source codes include source codes for multiple access networks such as the Slepian-Wolf network and Ahlswede-Körner network and source codes for the broadcast network such as the Gray-Wyner network. Examples of lossy network source codes also include source codes for multiple access networks and broadcast networks. Examples of broadcast network source codes include the multiresolution source code and multiple description source code. Examples of multiple access network source codes include the multiterminal source code and source codes for the CEO problem.

Algorithms

Lossless

• Huffman code
• Shannon code
• Shannon-Fano-Elias code
• Arithmetic code
• Tunstall code
• Burrows Wheeler transform

Lossy

• Vector quantization
• JPEG
• MPEG
• SPIHT

Universal

• Lempel-Ziv algorithm
• Context-tree weighting

Source Coding Limits, Bounds, and Inequalities

• Kraft's inequality
• Entropy
• Rate-distortion bound