Open AccessDetection of Glaucoma using ORB (Oriented FAST and Rotated BRIEF) Feature Extraction
Author(s) -
Kazi Safayet Md. Shabbir,
Md. Imteaz Ahmed,
Marzan Alam
Publication year - 2021
Publication title -
journal of engineering advancements
Language(s) - English
Resource type - Journals
eISSN - 2708-6437
pISSN - 2708-6429
DOI - 10.38032/jea.2021.03.005
Subject(s) - orb (optics) , artificial intelligence , glaucoma , computer science , computer vision , scale invariant feature transform , feature extraction , pixel , rotation (mathematics) , fundus (uterus) , image processing , pattern recognition (psychology) , image (mathematics) , ophthalmology , medicine
This research was utilized to identify glaucoma, a type of eye illness. This endeavor necessitates the use of pictures from the fundus camera for image processing. This study reflects the effort done to detect glaucoma-affected eyes utilizing image feature extraction using Oriented FAST and Rotated BRIEF (ORB). ORB is a binary descriptor approach that is based on BRIEF and is highly fast. This technique is insensitive to picture noise and is invariant to any rotation. ORB is two orders of magnitude faster than SURF and performs similarly to SIFT. It is more efficient than other texture analysis methods. It is less computationally difficult than other approaches in the literature. This technique extracts features and detects texture by inspecting each pixel of the retina picture. It was trained on 160 fundus pictures of normal and glaucoma-affected retinas. After that, any healthy or glaucoma-affected eye may be easily recognized by obtaining an accurate eye picture. The results reveal that this technique has a precision and accuracy of more than 90%.