Efficient software implementation of the JBIG compression standard

JBIG is a new binary image compression standard that is designed to handle both text and halftoned documents. It significantly outperforms the CCITT Group 3 and Group 4 standards especially on halftoned documents. The JBIG standard is based on an arithmetic encoder, and it features adaptive probability estimation, adaptive templates, and three different types of prediction. It also allows either sequential or progressive compression of image data. A brief discussion of information theory is given as it applies to image compression. The JBIG algorithm is described, and several techniques are developed to efficiently implement the algorithm in software. The most important techniques include an efficient scheme for building the context that are required, and taking advantage of large all-white or all-black regions of images by designing very efficient loops for processing those areas. Other techniques are also discussed, such as, efficient implementation of deterministic prediction, and an improved method for handling the conditional exchange condition. Timing information is given for the final implementation on several platforms. The JBIG algorithm is compared with the CCITT Group 3 and Group 4 algorithms, and it is tested in noisy environments