An efficient microaneurysms detection approach in retinal fundus images

Abstract

Diabetic retinopathy (DR) is one of the retinal disorders and the leading cause of blindness worldwide. Microaneurysms (MA) is the first clinical indication of DR, and the detection of MA helps in early diagnosis. The retinal fundus image analysis helps screen DR through MA detection. In general, the MA detection method consists of preprocessing, enhancement, and classification stages. Preprocessing is crucial to improve the retinal features and reduce the imaging artifacts. Reducing these artifacts is one of the challenging research problems in retinal fundus image analysis. In this paper, a novel improved Non-Local Mean filter (INLMF) is proposed to remove the imaging artifacts. The proposed method is tested on publicly available databases and images collected from Hospital. The proposed method has achieved the best performance metric than the state-of-the-art. The computational time per image is 6.2 sec which is less than other methods. © 2022, The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature.