Efficient reconfigurable architectures for 3D medical image compression

In this research, novel architectures based on different design approaches and arithmetic techniques such as direct mapping implementation, dynamic partial reconfiguration (DPR) mechanism, distributed arithmetic (DA) and systolic array (SA) will be developed for three dimensional (3D) medical image compression system. Moreover, solutions for processing large medical volumes will be investigated and power modelling of the architectures developed will be carried out on different field programmable gate array (FPGA) platforms. The ultimate aim of this research is to examine the most efficient reconfigurable architectures for 3D medical image compression. In this paper, the research framework and a case study addresses the performance of 3D Haar wavelet transform (HWT) with DPR mechanism are discussed and evaluated. Results obtained have shown the advantages offered by DPR and lead to a promising solution for implementing computationally intensive applications such as 3D medical image compression.