DCT IMAGE COMPRESSION USING SLIDING RLETECHNIQUE

Abstract

The discrete cosine transform (DCT) is a method for converting a signal into
plain frequency components. It is extensively used in image compression. In this
paper a new technique is proposed, namely SRLE (Sliding Run Length Encoding)
which is based on a lossy compression, and used to enhance image data
compression.
Image quality is measured impartially, using peak signal-to-noise ratio (PSNR) or
picture quality scale, and individually using perceived image quality with
compression factor (CF) being the main theme of this paper, taking into
consideration the preservation of well PSNR outputs.
The performance of DCT compression generally degrades low bit-rates mainly
because of the underlying block-based DCT scheme. Experimental results
demonstrate the effectiveness of the SRLE approach, in terms of PSNR, CF and
execution time, over different mechanisms used with DCT image compression. The
new technique also proved to have favourable results in terms of PSNR and CF
when compared with some wavelet based image compression.