Image Compression

Image compression involves reducing the size of image data files, while retaining necessary information. Files can be reduced in size by taking advantage of redundancies in images and the limitations of the human visual system. The important or necessary information to be retained is application specific. Compression algorithm development starts with applications to two-dimensional (2-D) still images and can be extended to video; this chapter focuses on compression for 2-D still images. A general model for compression is presented: (1) data reduction, (2) mapping, (3) quantization and (4) coding. Lossless algorithms, often required for medical or legal reasons, can only use the mapping and coding steps in the general model. Lossless compression methods discussed include Huffman coding, Golomb-Rice coding, run-length coding, Lempel-Ziv-Welch coding and arithmetic coding. Lossy methods discussed include gray-level run-length coding, block truncation coding, vector quantization, differential predictive coding, model-based and fractal compression, transform coding, hybrid and wavelet methods. Along with the text are 37 illustrative figures and 104 associated monochrome and color images. The end of chapter exercises includes problems and programming exercises.