Wednesday, December 24, 2008

Collaborating Text and Image Features for Building Decision Tree of Diagnose Problem

Making a good and clear diagnosis needs medical data in various forms. A design of diagnosis application is discuss here using case based reasoning method. All previous domain cases are represented in decision tree, text and medical image visual features data concerning with the domain are used as variables decision.

The texture features extracted from medical images used here are first order color moment, second order color moment, third order color moment, entropy, energy, contras and homogeneity. The decision tree is constructed by C4.5 algorithm.

The build decision tree hopefully can support diagnosis process in computer based diagnosis tool.


....

to read the full paper you can download here

this paper has been presented in International Graduate Conference on Engineering and Science on 24th December 2008 in University Technology Malaysia

the presentation file can be download here