Color and texture analysis of cervix lesions

Date

2004-12

Journal Title

Journal ISSN

Volume Title

Publisher

Texas Tech University

Abstract

This work focuses on selection of appropriate geometrical and color features for classification and morphological texture analysis for characterization of diagnostically significant vascular patterns embedded in cervical lesions observed in Colposcopic and Cervicographic images of the ectocervix. Morphological texture analysis included a sequence of operations to yield skeletonized vascular structures uniquely representing cancerous and non-cancerous tissue stractures which are not easily observed in Colposcopic and Cervicographic images. This algorithmic ability to classify and characterize the embedded vascular patterns in cervical lesions may lead to automated and precise diagnostic differentiation of underlying vascular structure.

Description

Keywords

Diagnostic imaging, Cervix uteri -- Tumors -- Imaging, Imaging systems in medicine, Colposcopy, Tumors -- Classification

Citation