Abstract
In this paper, we present our preliminary results classifying benign and malignant tissues within the prostate peripheral zone using textons. For this purpose, patches are randomly extracted from malignant and benign regions and we perform k-means clustering to generate textons. All textons are combined to form the texton dictionary which was used to construct a texton map for every peripheral zone region in each training image. Based on the texton map, histogram models for each malignant and benign tissue are constructed which will be used to train our classifiers. We tested the proposed method on 418 T2-W MR images taken from 45 patients and evaluation results of Az = 87% ±7% show a comparable performance with the state-of-the-art in the literature.
| Original language | English |
|---|---|
| Pages | 138-143 |
| Number of pages | 6 |
| Publication status | Published - 15 Jul 2015 |
| Event | 19th Conference on Medical Image Understanding and Analysis 2015 - Lincoln, United Kingdom of Great Britain and Northern Ireland Duration: 15 Jul 2015 → 17 Jul 2015 |
Conference
| Conference | 19th Conference on Medical Image Understanding and Analysis 2015 |
|---|---|
| Abbreviated title | MIUA |
| Country/Territory | United Kingdom of Great Britain and Northern Ireland |
| City | Lincoln |
| Period | 15 Jul 2015 → 17 Jul 2015 |