Certain investigations on edge directed interpolation based image compression for multimedia applications

Abstract

Recent technological developments in image compression have been drastically influenced by the field of wavelet transform The wavelet transform plays a vital role in many image compression algorithms The main objective of image compression is to reduce redundancy and irrelevancy of image data The image compression is inevitable in multimedia communications The data handling in multimedia is ever increasing nowadays which needs efficient transmission and memory to store it Basically there are two elementary types of image compression namely lossless and lossy image compression In lossless image compression the reconstructed image is same as original The application of lossless compression is more suitable where the loss of information becomesintolerable Medical imaging and text compression are the best example of lossless image compression which gives minimal compression ratio On the other hand lossy compression which produces high compression ratio and the reconstructed image is close approximation of the original image The advantage of lossy compression is high coding speed and plenty of applications where the loss of information is unnoticeable by HVS system. Discrete Wavelet Transform DWT and Set Partitioning In Hierarchical Trees (SPIHT) are the two key enabling techniques in image compression DWT is well standard algorithm in the field of still image compression. It is more popular and it has desirable properties such as multiresolution analysis vanishing moments energy compaction etc SPIHT is rate scalable, high coding performance and operated on the coefficients of DWT that are formed as a Spatial Orientation Tree SOT structure SPIHT has much lower computational complexity fully embedded coded file with progressive coding techniques simple algorithm and exact bit rate coding and also applicable to lossless image compression Thus evolves interest to improve the performance of SPIHT algorithm newline