Image Compression Using Run Length and Variable Length Coding

Abstract

A revolution in information and entertainment has been widespread across the world. Information contents in digital form are pulsing through cables, telecommunication networks and direct satellites delivery system viewed on PC, stored on RAM disk, tape, CD-Rom etc. The digital information contains text, audio and video. Among these, audio and video pose real challenges because both of these require real time operation and produce large data files when they are digitized. To overcome these problems either cables with high bandwidth can be used or data compression technique can be used. Compression of data is a better alternative. A foregoing research and many implementations using software, hardware or both for, two algorithms, one for compression of data at source encoding and another for decompression at its destination decoding, has established that transformation coding achieves larger compression than predictive coding. In transformation coding compression is achieved by transforming the given image into another array such that a large amount of information ios packed into small number of samples. While predictive coding involves predicting subsequent values by observing previous ones and then transmitting only the small difference between actual and predicted data. Any distortion due to quantization and channel errors gets distributed during inverse transformations. Predictive coding is quite sensitive to changes in the statistics of the data and in two dimensions, finite order casual predictors may neverf achieve compression ability close to transform coding. Discrete Cosine Transformations. The choice of coding is situation dependent. In general a single coding is employed that may be variable length coding or run length coding. This gives a satisfactory compression ratio up to 50 but still is not very good. The present work employs the combination of both the coding techniques i.e. image is first coded as run length of levels and then run lengths are again coded using variable codes. The outcome of this work is a software system for gray scale image compression which is based on loss compression. This technique has the advantage of providing better compression ratio i.e. 10 to 95 times over the other existing techniques which compress images up to 50 times. Further this technique is very flexible as it gives user the choice for selecting compression ratio and retrieved image quality.