Skip to Main Content
Detailed process analysis and different optimization strategies are necessary in order to make wavelet based image processing algorithms usable for portable multimedia devices. Besides optimizing the hardware structures for better performance, the overall processing load has to be minimized as proposed in this paper. Analyzing the image content adoptively for all decomposition levels and subbands, signal entropy can be reduced to obtain shorter code words after the entropy coder. We propose a reduction of the data accuracy in the quantizer for high compression rates, saving up to 28% processing power without compromising the image quality.