Quantizer Effects on EBCOT Based Compression
Christian Guenter, Albrecht Rothermel · 2007
In order to make wavelet based image processing algorithms usable for portable multimedia devices different optimization strategies need to be taken into account. Besides optimizing the hardware structures for better performance, the overall processing load has to be minimized as proposed in this paper. Coarse quantizer steps sizes reduce the entropy and therefore we get shorter code words after the entropy coder. We propose a reduction of the data accuracy in the quantizer, saving up to 28% processing power. Image quality and compression ratio are not compromised.