ABSTRACT
Data from various sources often have non-uniform probability distributions. Lossless coding achieves data compression by taking advantage of the non-uniform probability distribution of the data (e.g., using shorter code-words for more frequent symbols and longer code-words for less frequent symbols). It is widely used in text, facsimile, image, and video compression [1]–[2]. It can be used alone (e.g., in text, facsimile, and medical image compression) or used with a lossy coder (e.g., in most image/video compression). Most image/video coders consist of a lossy coder to remove the spatial and temporal redundancies, and a lossless coder to remove the statistical redundancies. With a lossless coder, an image/video coder can achieve an additional compression ratio of about two beyond what can be achieved with a lossy coder alone.
