It is well known that retinal diseases are sometimes identified by tortuosity of the vessels, presence of exudates and hemorrhages while lesions of tissues are associated to diabetic retinopathy, retinopathy of prematurity and more general cerebrovascular problems. One of the main issues in this research field is detecting small curvilinear structures, thus the aim of this contribution is to introduce a non-supervised and automated methodology to detect features such as curvilinear structures in retinal images. The core of the proposed methodology consists in using an approach that resembles the 'à trous' wavelet algorithm. With respect to the standard Gabor analysis our methodology is based on a sequence Gaussian filters, it is faster yet effective in the representation of the directions along the retinal vessels, which is a useful information to evaluate their tortuosity and to segment the images. To evaluate the correctness of the results we carried out a comparison with the so called Scale and Curvature Invariant Ridge Detector, which is considered as one of the most effective supervised methods for retinal vessel detection, on a pair of public domain datasets

Lo Castro, D., Tegolo, D., Valenti, C. (2018). Filter Bank: a Directional Approach for Retinal Vessel Segmentation. In S. Qiu, H. Liu, L.W. Sun, Q. Li, M. Zhou (a cura di), Proceedings - 2017 10th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics, CISP-BMEI 2017 (pp. 1-6). Institute of Electrical and Electronics Engineers Inc. [10.1109/CISP-BMEI.2017.8302192].

Filter Bank: a Directional Approach for Retinal Vessel Segmentation

Lo Castro, Dario;Tegolo, Domenico;Valenti, Cesare
2018

Abstract

It is well known that retinal diseases are sometimes identified by tortuosity of the vessels, presence of exudates and hemorrhages while lesions of tissues are associated to diabetic retinopathy, retinopathy of prematurity and more general cerebrovascular problems. One of the main issues in this research field is detecting small curvilinear structures, thus the aim of this contribution is to introduce a non-supervised and automated methodology to detect features such as curvilinear structures in retinal images. The core of the proposed methodology consists in using an approach that resembles the 'à trous' wavelet algorithm. With respect to the standard Gabor analysis our methodology is based on a sequence Gaussian filters, it is faster yet effective in the representation of the directions along the retinal vessels, which is a useful information to evaluate their tortuosity and to segment the images. To evaluate the correctness of the results we carried out a comparison with the so called Scale and Curvature Invariant Ridge Detector, which is considered as one of the most effective supervised methods for retinal vessel detection, on a pair of public domain datasets
Settore INF/01 - Informatica
9781538619377
https://ieeexplore.ieee.org/document/8302192/
Lo Castro, D., Tegolo, D., Valenti, C. (2018). Filter Bank: a Directional Approach for Retinal Vessel Segmentation. In S. Qiu, H. Liu, L.W. Sun, Q. Li, M. Zhou (a cura di), Proceedings - 2017 10th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics, CISP-BMEI 2017 (pp. 1-6). Institute of Electrical and Electronics Engineers Inc. [10.1109/CISP-BMEI.2017.8302192].
File in questo prodotto:
File Dimensione Formato  
Filter Bank - a Directional Approach for Retinal Vessel Segmentation.pdf

Solo gestori archvio

Tipologia: Versione Editoriale
Dimensione 6.89 MB
Formato Adobe PDF
6.89 MB Adobe PDF   Visualizza/Apri   Richiedi una copia

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/10447/290971
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 5
  • ???jsp.display-item.citation.isi??? 0
social impact