A DWT-DCT image compression scheme

The heavy growth of multimedia based web applications has increased the need for efficient coding of signals and images and has made compression of such signals central to storage and communication technology. The block based Discrete Cosine Transform (DCT) has been the building block of most image compression algorithms and has been adopted in the earlier JPEG standard and in the MPEG series video standards. Over the last decade, research activities in image coding have centered largely around the Discrete Wavelet Transform (DWT) which provides an elegant representation suitable for compression and as such it has been included in the latest JPEG-2000 standard. With proper re-ordering of the DCT coefficients, efficient quantization schemes developed in the context of wavelet compression can be applied in the DCT domain. The objective of this thesis is to combine the better features of both the transformations and develop a hybrid image coding algorithm with low computational complexity. An image is decomposed using a one-level DWT and then a block based DCT is applied to the low-pass subband. The quantization scheme adopted is the Difference Reduction algorithm (DR) to obtain an embedded, progressive image compression scheme. The scheme performs better than the DCT based coders and shows good subjective performance over a wide range of compression ratios.