Retinal Disease Screening Through Local Binary Patterns

Abstract

Morales, S.; Engan, K.; Naranjo Ornedo, V.; Colomer, A. (2015). Abstract-This work investigates discrimination capabilities in the texture of fundus images to differentiate between pathological and healthy images. For this purpose, the performance of Local Binary Patterns (LBP) as a texture descriptor for retinal images has been explored and compared with other descriptors such as LBP filtering (LBPF) and local phase quantization (LPQ). The goal is to distinguish between diabetic retinopathy (DR), agerelated macular degeneration (AMD) and normal fundus images analysing the texture of the retina background and avoiding a previous lesion segmentation stage. Five experiments (separating DR from normal, AMD from normal, pathological from normal, DR from AMD and the three different classes) were designed and validated with the proposed procedure obtaining promising results. For each experiment, several classifiers were tested. An average sensitivity and specificity higher than 0.86 in all the cases and almost of 1 and 0.99, respectively, for AMD detection were achieved. These results suggest that the method presented in this paper is a robust algorithm for describing retina texture and can be useful in a diagnosis aid system for retinal disease screening.

An Effective Diagnosis of Diabetic Retinopathy with Aid of Soft Computing Approaches

“…Local binary patterns are well known method for screening retinal disease explained in Ref. [8]. Automatic segmentation of blood vessels is in Refs.…”

A Survey on Intelligent Screening for Diabetic Retinopathy

“…For example, Agurto et al7 extracted texture features by applying multiscale amplitude-modulation-frequency-modulation. In the study of Morales et al,76 the local binary patterns were used as a texture descriptor to distinguish DR from normal. Mixed features including area of EXs and vessels, texture and entropy, were used for DR classification by Mookiah et al,77 and three classifiers were tested.…”

Fundus image analysis for automatic screening of ophthalmic pathologies.

“…During those years, a lot of works related to the automatic detection of DR and AMD have been published (Colomer et al, 2017a(Colomer et al, ,b, 2018Morales et al, 2017Morales et al, , 2013Morales et al, , 2014. In this way, the research project entitled "Fundus image processing for automatic screening of ophtalmological diseases -ACRIMA" (TIN2013-46751-R) was funded by the Ministerio de Economía y Competitividad of Spain.…”

“…In the particular case of fundus images, LBP technique has mostly been used on the automatic retinal vessel detection (Hatami and Goldbaum, 2016;Moin et al, 2010;Zabihi et al, 2010) rather than on the identification and localisation of abnormal retinal tissue. However, some state-of-the-art works focused on the automatic detection of DR and AMD can be found (Garnier et al, 2014;Mookiah et al, 2013;Morales et al, 2017). Garnier et al (2014) extract the texture information at serveral scales through wavelet decomposition and LBP with the aim of detecting AMD.…”

“…The work of Morales et al (2017) (in which A. Colomer significantly contributed) was the starting point of this thesis chapter. In that work, the goal was to distinguish between DR, AMD and normal fundus images avoiding any previous segmentation stage of retinal lesions.…”