An FPGA based coprocessor for the classification of tissue patterns in prostatic cancer

Abstract

This paper discusses the suitability of reconfigurable computing to speedup medical image classification problems. As an example of the speedup offered by reconfigurable logic, a multispectral computer vision system for automatic diagnosis of prostatic cancer is implemented. Different parallel architectures for various steps in automatic diagnosis are proposed and implemented in Field Programmable Gate Arrays (FPGAs). The first step of the algorithm is to compute Grey Level Cooccurrence Matrix (GLCM). The second step involves the normalisation of GLCM. The third step of the algorithm is to compute texture features from the normalised GLCM. The last step is concerned with image classification using linear discriminant analysis (LDA). Finally, the performance of the proposed system is assessed and compared against a microprocessor based solution. The results obtained clearly show that the proposed solution compares favorably.