RT Generic
T1 Computer aided diagnosis system using dermatoscopical image
A1 López Labraca, Javier
AB Computer Aided Diagnosis (CAD) systems for melanoma detection aim to mirror the expertdermatologist decision when watching a dermoscopic or clinical image. Computer Visiontechniques, which can be based on expert knowledge or not, are used to characterize thelesion image. This information is delivered to a machine learning algorithm, which gives adiagnosis suggestion as an output.This research is included into this field, and addresses the objective of implementing acomplete CAD system using ‘state of the art’ descriptors and dermoscopy images as input.Some of them are based on expert knowledge and others are typical in a wide variety ofproblems. Images are initially transformed into oRGB, a perceptual color space, looking forboth enhancing the information that images provide and giving human perception to machinealgorithms. Feature selection is also performed to find features that really contribute todiscriminate between benign and malignant pigmented skin lesions (PSL). The problem ofrobust model fitting versus statistically significant system evaluation is critical when workingwith small datasets, which is indeed the case. This topic is not generally considered in worksrelated to PSLs. Consequently, a method that optimizes the compromise between these twogoals is proposed, giving non-overfitted models and statistically significant measures ofperformance. In this manner, different systems can be compared in a fairer way. A databasewhich enjoys wide international acceptance among dermatologists is used for theexperiments.
YR 2014
FD 2014
LK https://hdl.handle.net/10016/26651
UL https://hdl.handle.net/10016/26651
LA eng
DS e-Archivo
RD 30 abr. 2024