Application of soft computing to medical image processing, analysis and computer-aided diagnosis

Du-Yih Tsai

Department of Radiological Technology, School of Health Sciences, Niigata University, Japan

Soft computing is a collection of methodologies that exploit a tolerance for imprecision, uncertainty and partial truth. Results are achieved with tractability and robustness. These results are often better than that what is achieved through the exclusive use of conventional (hard) computing. The constituents of soft computing include fuzzy logic, artificial neural networks, genetic algorithms, evolutionary computing, chaos, etc.

In recent years, medical imaging technologies, such as image processing and analysis, have evolved at an explosive rate. Medical image processing and analysis is concerned with the enhanced visualization, combined use and quantification of medical images for optimal diagnosis, planning of therapy and therapy follow-up. In the context of medical imaging, soft computing technique appears as a power framework since it can lead successful tasks such as segmentation, interpretation, classification and so on.

This talk describes two major approaches to medical image segmentation and classification using soft computing technology. One uses an artificial neural network to segment liver structure in CT images. The other uses a genetic-algorithm-based fuzzy-logic approach for computer-aided diagnosis scheme in medical imaging. The scheme is applied to discriminate myocardial heart disease from echocardiographic images and to detect and classify clustered microcalcifications from mammograms.