INTRODUCTION This research describes an image representation technique that entails progressive refinement of user specified regions of interest (ROI) of large images. Progressive refinement of original quality can be accomplished in theory. However, due to heavy burden on storage resources for our applications, we restrict the refinement to about 25% of the original data resolution. Wavelet decomposition with Vector Quantization (VQ) of the high frequency components and JPEG/DCT compression of low frequency component is used …showed first 75 words of 533 total…

…showed last 75 words of 533 total…IMAGE DECOMPRESSION The reconstruction program loads the LL subband by decompressing the Low-Low wavelet component using JPEG/IDCT. The reconstruction program then fetches the MVQ indices of the corresponding ROI and reconstructs a better quality image. A three level block diagram of reconstruction is shown in Figure 6. The reconstruction can be iterated on user’s request until sufficient image quality is displayed for the user to make decision about placing an order for the image.