Retinal Fundus Image Supported Examination of Macular Scar using Handcrafted Features

Eye is a vital organ in human physiology and its illness will cause various vision issues. Early diagnosis and management of eye health is essential to treat the vision issues. The clinical approaches include the personal check and image based eye examination. Retinal Fundus Image (RFI) is a common imaging modality for analyzing the eye illness and this work considered the examination of Macular Scar (MS) for the study. This study proposed a machine-learning tool to examine the healthy/MS RFI-database and to achieve better efficiency, it implemented Spider-Wasp algorithm based features optimization. The different phases in the ML-tool includes; image collection and initial adjustment, features extraction using local binary pattern and discrete wavelet transform, features reduction with considered scheme, serial features fusion to generate fused-features vector, and classification and 3-fold cross validation to confirm the performance. This research implemented a binary-classification and the experimental outcome presents an accuracy of >92% during the healthy/MS classification.