5.1 - Lossless Coding

Abstract

The goal of lossless image compression is to represent an image signal with the smallest possible number of bits without a loss of any information, thereby speeding up transmission and minimizing storage requirements. This chapter introduces the basics of lossless image coding and presents classical as well as some more recently developed lossless compression methods. Lossless symbol coding is commonly referred to as “lossless coding” or “lossless compression.” The popular lossless symbol coding schemes fall into one of the following main categories: statistical schemes (Huffman, Arithmetic) and dictionary-based schemes (Lempel-Ziv). Statistical schemes require the knowledge of the source symbol probability distribution. Dictionary-based schemes do not require a priori knowledge of the source symbol probability distribution. They dynamically construct encoding and decoding tables of variable-length symbol strings as and when they occur in the input data. The operations of a lossless image encoder can be grouped into three stages: transformation, data to symbol mapping, and lossless symbol coding. The factors that need to be considered when choosing or devising a lossless compression scheme are compression efficiency, coding delay, implementation complexity, robustness, and scalability.