Tumor detection using iris pattern
Cancer is one of the top killer diseases in the world.
A computerized system or an expert system of the iridology, which could diagnose cancer using patient iris image. The system could be split into 3 sections which are locating region of interest, extracting data and analyzed data. The raw iris image was pre-process
using Hough Transfrom to extract the iris region. Histogram threshold method and color model was applied to the image to segment the cancer feature. A rule base algorithm was used to analyze the cancer features and determine the location of the cancer. The iris region was successfully located and extracted from out sample. 70% of the cancer feature was successfully identified by our method. Very accurate or full system to diagnose cancer patient is still under research. More cancer patient and potential cancer patient needed to improve the accuracy of the system.
Lye Wil Liam, Ali Chekima, Liau Chung Fan, Jamal Ahmad Dargham
School of Engineering and Information Technology, University Malaysia Sabah, Beg Berkunci 2073,
88999 Kota Kinabalu, Sabah Malaysia.
Reference: http://web.usm.my/nsf/abstracts/aICT01.pdf
Leave a Reply