Real-world medical applications constitute a wide research area in computer sciences. Human factor in the current practice of medicine, such as cancer diagnosis, causes subjectivity. The aim of computer sciences is to reduce subjectivity and bring objective diagnosis, affecting the selection of treatment type. In our research work, our goal is to introduce new statistical diagnosis of cancer that makes use of quantitative measures to objectively diagnose the cancer, and thus improve the diagnosis accuracy. We make use of the constraint voronoi diagrams (VD) and delaunay triangulations (DT) for the representation of tissues and propose to use the features extracted from these constraint VDs and DTs. In this talk, we first discuss a general framework for VDs and DTs, how to constrain VDs and DTs for our purpose, and then how to define metrics. Finally, we present our preliminary results.

DATE: 26 November, 2007, Monday@ 16:15
PLACE: EA 409