Diagnostics 2022, 12, 1344. https://doi.org/10.3390/diagnostics12061344 www.mdpi.com/journal/diagnostics Article Intelligent Diagnosis and Classification of Keratitis Hiam Alquran 1,2 , Yazan Al-Issa 3 , Mohammed Alsalatie 4 , Wan Azani Mustafa 5,6, *, Isam Abu Qasmieh 2 and Ala’a Zyout 2 1 Department of Biomedical Engineering, Jordan University of Science and Technology, Irbid 22110, Jordan; heyam.q@yu.edu.jo 2 Biomedical Systems and Medical Informatics Engineering, Yarmouk University, Irbid 21163, Jordan; iabuqasmieh@yu.edu.jo (I.A.Q.); alzuet@yu.edu.jo (A.Z.) 3 Department of Computer Engineering, Yarmouk University, Irbid 21163, Jordan; alissay@yu.edu.jo 4 The Institute of Biomedical Technology, King Hussein Medical Center, Royal Jordanian Medical Service, Amman 11855, Jordan; mhmdsliti312@gmail.com 5 Faculty of Electrical Engineering Technology, Campus Pauh Putra, Universiti Malaysia Perlis, Arau 02600, Malaysia 6 Advanced Computing (AdvComp), Centre of Excellence (CoE), Campus Pauh Putra, Universiti Malaysia Perlis (UniMAP), Arau 02600, Malaysia * Correspondence: wanazani@unimap.edu.my Abstract: A corneal ulcer is an open sore that forms on the cornea; it is usually caused by an infection or injury and can result in ocular morbidity. Early detection and discrimination between different ulcer diseases reduces the chances of visual disability. Traditional clinical methods that use slit- lamp images can be tiresome, expensive, and time-consuming. Instead, this paper proposes a deep learning approach to diagnose corneal ulcers, enabling better, improved treatment. This paper sug- gests two modes to classify corneal images using manual and automatic deep learning feature ex- traction. Different dimensionality reduction techniques are utilized to uncover the most significant features that give the best results. Experimental results show that manual and automatic feature extraction techniques succeeded in discriminating ulcers from a general grading perspective, with ~93% accuracy using the 30 most significant features extracted using various dimensionality reduc- tion techniques. On the other hand, automatic deep learning feature extraction discriminated sever- ity grading with a higher accuracy than type grading regardless of the number of features used. To the best of our knowledge, this is the first report to ever attempt to distinguish corneal ulcers based on their grade grading, type grading, ulcer shape, and distribution. Identifying corneal ulcers at an early stage is a preventive measure that reduces aggravation and helps track the efficacy of adapted medical treatment, improving the general public health in remote, underserved areas. Keywords: corneal ulcer; deep learning; ResNet101; PCA 1. Introduction Corneal ulcer (CU), also known as keratitis, is an infection or inflammation that af- fects the transparent anterior portion of the eye that covers the iris, which is known as the cornea [1]. Corneal ulcer is a major cause of sight loss and might be responsible for 1.5– 2.0 million blindness cases every year [2]. The source of corneal ulcer can be viral, bacte- rial, fungal, or parasitic. The symptoms include pain, ache, redness, blurry vision, and sensitivity to bright light. Traditional methods that use slit-lamp images and slit-lamp microscopy for diagnosing corneal ulcers are subjective and time-consuming, and they are highly dependent on the ophthalmologist expertise. It is a preventable and treatable disease, as early and timely recognition of corneal ulcer can stop the deterioration and help maintain a patient’s visual integrity. Developments in staining techniques help investigators numerically detect and di- agnose ulcers. Fluorescein is a widely used dye in ophthalmology for the diagnosis and Citation: Alquran, H.; Al-Issa, Y.; Alsalatie, M.; Mustafa, W.A.; Qasmieh, I.A.; Zyout, A. Diagnosis and Classification of Keratitis Using Artificial Intelligence Approach. Diagnostics 2022, 12, 1344. https:// doi.org/10.3390/diagnostics12061344 Academic Editors: Jae-Ho Han and Christoph Palm Received: 23 April 2022 Accepted: 26 May 2022 Published: 28 May 2022 Publisher’s Note: MDPI stays neu- tral with regard to jurisdictional claims in published maps and institu- tional affiliations. Copyright: © 2022 by the authors. Li- censee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and con- ditions of the Creative Commons At- tribution (CC BY) license (https://cre- ativecommons.org/licenses/by/4.0/).