Title :
Unified methodology for evaluating vessel tree tortuosity metrics in eye fundus images
Author :
Sanchez, L. ; Novo, J. ; Fernandez, Alicia ; Barja, J.M.
Author_Institution :
Dept. of Comput. Sci., Univ. of A Coruna, A Coruna, Spain
Abstract :
The tortuosity of a vessel, that is, how and how many times it bends, is a relevant measure to diagnose different pathologies. However, there is an important drawback, as each specialist has its own estimation about what a tortuous vessel is, and the process has to be evaluated manually, implying a large and tedious work. Numerous studies had attempted to define a reliable and deterministic measure for this parameter, but employing different approaches, inputs and validation criteria. In this article, we intended to implement a unified methodology that integrates the different metrics and also to establish a comparison among them, employing the same inputs, validation criteria and procedure for the vascular tree extraction.
Keywords :
biomedical optical imaging; blood vessels; eye; bending; eye fundus image; pathology diagnosis; pathology measurement; unified methodology; vascular tree extraction; vessel tree tortuosity metric evaluation;
Conference_Titel :
Computer-Based Medical Systems (CBMS), 2013 IEEE 26th International Symposium on
Conference_Location :
Porto
DOI :
10.1109/CBMS.2013.6627862