Tumor Detection In Colonoscopy (2004)
| Venue: | In Fourth European Symposium on “Biomedical Engineering |
BibTeX
@INPROCEEDINGS{Unsupervised04tumordetection,
author = {Using The Unsupervised and G. D. Magoulas and V. P. Plagianakos and D. K. Tasoulis and M. N. Vrahatis},
title = {Tumor Detection In Colonoscopy},
booktitle = {In Fourth European Symposium on “Biomedical Engineering},
year = {2004},
pages = {25--27}
}
Bookmark
 |
 |
 |
 |
 |
 |
OpenURL
Abstract
This study presents an approach to automatically detect tumors in colonoscopic images that is based on the synergy between unsupervised clustering and artificial neural networks. First the noisy data set is partitioned into clusters and then a di#erent neural network is trained from data of each detected cluster. Each network is therefore considered a "local expert" with regards to patterns of the same cluster as the one used for training. The results from applying this approach in interpreting colonoscopy images are promising, generating networks that were able to detect malignant regions of interest with high accuracy.
