Automated detection of circinate exudates in retina digital images using empirical mode decomposition and the entropy and uniformity of the intrinsic mode functions

Lahmiri, Salim et Boukadoum, Mounir (2014). « Automated detection of circinate exudates in retina digital images using empirical mode decomposition and the entropy and uniformity of the intrinsic mode functions ». Biomedical Engineering / Biomedizinische Technik, 59(4), pp. 357-366.

Fichier(s) associé(s) à ce document :

Prévisualisation	PDF Télécharger (3MB)

Résumé

This work presents a new automated system to detect circinate exudates in retina digital images. It operates as follows: the true color image is converted to gray levels, and contrast-limited adaptive histogram equalization (CLAHE) is applied to it before undergoing empirical mode decomposition (EMD) as intrinsic mode functions (IMFs). The entropies and uniformities of the first two IMFs are then computed to form a feature vector that is fed to a support vector machine (SVM) for classification. The experimental results using a set of 45 images (23 normal images and 22 images with circinate exudates taken from the STARE database) and tenfold cross-validation indicate that the proposed approach outperforms previous works found in the literature, with perfect classification. In addition, the image processing time was < 4 min, making the presented circinate exudate detection system fit for use in a clinical environment.

Type:	Article de revue scientifique
Mots-clés ou Sujets:	automated detection; circinate exudates; classification; empirical mode decomposition; feature extraction; retina.
Unité d'appartenance:	Faculté des sciences > Département d'informatique
Déposé par:	A. Mounir Boukadoum
Date de dépôt:	15 avr. 2016 13:04
Dernière modification:	27 avr. 2016 18:10
Adresse URL :	http://archipel.uqam.ca/id/eprint/8144

Modifier les métadonnées (propriétaire du document)

Statistiques