Data Mining Technique for Biomedical Image Classification

Prof. Jong-Min Park
 

Department of Electrical and Computer Engineering, San Diego State University

2002. 5. 29

 
This presentation describes a data mining framework that aids in the process of finding an optimal set of features and its application into classification and detection of glaucoma disease from optic nerve data.
An overview of data mining and machine learning techniques, and their application to biomedical image classification mechanism will be introduced. Examples of early glaucoma detection systems using machine learning will be shown.
There are problems and issues when applying data mining to biomedical classification, including selection and evaluation of high-dimensional features from the optic nerve data, and several generations of optical systems.
Solutions to these issues using novel algorithms, the experimental results, and the implementation will be presented. The presentation will conclude with future works.


This page is maintained by Ho-Jin Chung (
hjchung@bi.snu.ac.kr).
Last update: May. 28, 2002