Trainable WEKA Segmentation of Retinal Fundus Images for Global Eye Disease Diagnosis Application
T Jebaseeli, C Anand, Deva Durai, J Peter, Donghuan Lu, Morgan Heisler, & H Nur, Tjandrasa, Muhammad Arsalan, Muhammad Owais, S Wang, Y Yin, G Cao, B Wei, Y Zheng, G Yang, R Vega, G Sanchez-Ante, L Falcon-Morales, H Sossa, E Guevara, Feng Li, & Zheng Liu, Lily Varun Gulshan, Peng, C Mahiba, A Jayachandran, Filippo Arcadu, Fethallah Benmansour, R Karthikeyan, P Alli, P Anilkumarb, Kumar, Ashutosh Kumar, Sazak, C Nelson, B Obara, M Alam, D Le, J Lim, R Chan, X Yao, Kranthi Kumar, Palatalise, Bhavani Sambaturu, Shailesh Kumar, Basant Kumar, B Wu, W Zhu, F Shi, S Zhu, X Chen, A Christodoulides, T Hurtut, H Tahar, F Cheriet, Rajesh Babu, K, T Manjula2, Thulasi Sivachandar Kasiviswanathan, Lorenzo Bai Vijayan, Giovanni Simone, Dimauro, Pedro Costa, Adrian Galdran, Asim Smailagic
Abstract
Image segmentation is a challenging task and quite important in biomedical images for disease identification and measuring the progress of the disease.Segmentation and extraction of retinal blood vessels, micro aneurysms, hemorrhages, soft exudates and hard exudates have a significant attention to diagnosis various retinal related diseases especially diabetic retinopathy and glaucoma.The proposed work effectively segments the above-mentioned regions from the retinal images with Contrast Limited Adaptive Histogram Equalization (CLAHE) preprocessing and Fast Random Forest Classifier supported by Trainable WEKA segmentation of FIJI Tool.The proposed work segments each and every pixel of the retinal image into predefined classes and the produced good segmentation results as shown in the paper.