An architecture-based framework for enhancing medical image compression using dyadic wavelet filters and context adaptive encoding with matrix-based post-processing

Medical image compression is popular due to increasing memory storage needs, promoting effective transmission and guaranteeing the preservation of diagnostic data without sacrificing patient care. Image compression is defined as the process of reducing the number of bits without losing the quality and information in an image. In medical image compression methods, the image has to be reproduced without losing its data. Despite various compression standards available, compression of images with better output quality and memory required for the compression process is still challenging. This paper introduces an optimized dyadic db3 orthogonal wavelet filter for reduced memory requirement purposes. In addition, this wavelet mechanism is integrated with matrix-based post-processing to reduce spatial redundancies and shows high compression performance. The hardware implementation of high-throughput context binary arithmetic coding is carried out for encoding purposes. The proposed system is implemented on a Xilinx working platform and implemented on an FPGA board. The proposed approach is implemented with a medical image dataset, and its performance is analyzed based on its throughput and frequency. Moreover, the quality of the input image is analyzed using the parameter compression ratio, peak signal to noise ratio, and maximum absolute error values. The proposed method is compared to conventional methods to demonstrate its superiority.