Abstract
In this work, a method for image analysis and processing and its application in natural and real medical images is presented. The Contourlet Transform (CT) was used to extract statistical texture features from texture images and Support Vector Machines (SVMs) were utilized for the classification process. A variety of filter and parameter combinations were tested in the experimental procedure in order to assess the efficiency of the method. This image analysis and processing procedure was applied to natural images taken from the publicly available Vistex library, and to real thyroid ultrasound (US) images. The evaluation of experimental results demonstrate the CT’s potential in feature extraction and shows that CT based texture features can be successfully utilized for texture representation.
Original language | English |
---|---|
Title of host publication | 1st International Conference for undergraduate and graduate students in Informatics and Related Applications |
Subtitle of host publication | EUREKA! 2010 |
Number of pages | 10 |
Publication status | Published - 1 Oct 2010 |
Event | 1st International Conference for Undergraduate and Postgraduate students in Computer Engineering, Informatics, related technologies and Applications - Patras - Ancient Olympia, Greece Duration: 15 Oct 2010 → 16 Oct 2010 http://www.uom.gr/modules.php?op=modload&name=News&file=article&tmima=1&categorymenu=7&sid=3465 |
Conference
Conference | 1st International Conference for Undergraduate and Postgraduate students in Computer Engineering, Informatics, related technologies and Applications |
---|---|
Abbreviated title | EUREKA! 2010 |
Country/Territory | Greece |
City | Patras - Ancient Olympia |
Period | 15/10/10 → 16/10/10 |
Internet address |
Keywords
- contourlet transform
- feature extraction
- texture analysis
- ultrasound images
- htyroid
- support vector machines