A PARALLEL SYSTEM FOR THE CLASSIFICATION OF CANCEROUS AND NORMAL COLONIC MUCOSA TISSUE IMAGES

摘要

Analysis of tissue using image processing techniques is useful for dealing with a number of problems in cancer research. Ideally in the future it will be possible to construct a fully automated computer system, one that can perform image classification without requiring human intervention. The aim of this research is to develop a system for performing classification of cancerous, dysplastic or normal colonic mucosa tissue images, by means of identifying the image processing techniques required, and experimenting with various classification techniques. A number of co-occurrence matrix feature extraction algorithms have been selected and are presented in this paper. Since analysis of tissue images is a complex task requiring vast processing power, parallel computing techniques have been employed. The classification system was implemented by means of a C++ library for distributed system programming using PVM (Parallel Virtual Machine) on a cluster of workstations. The performance and accuracy of the system are discussed in this paper.