Indonesian Journal of Electrical Engineering and Computer Science Vol. 28, No. 1, October 2022, pp. 209~219 ISSN: 2502-4752, DOI: 10.11591/ijeecs.v28.i1.pp209-219  209 Journal homepage: http://ijeecs.iaescore.com Enhancement of automatic classification of arcus senilis- nonarcus senilis using convolutional neural network Nur Farahin Bt Abdul Halim 1 , Ridza Azri Bin Ramlee 1,3 , Mohd Zaki Bin Mas’ud 2 , Amirul Jamaludin 1 1 Faculty of Electronics and Computer Engineering, Universiti Teknikal Malaysia Melaka (UTeM), Durian Tunggal, Melaka, Malaysia 2 Faculty of Information and Communication Technology, UTeM, Durian Tunggal, Melaka, Malaysia 3 Advanced Sensors Embedded Control System Research Group (ASECs), UTeM, Durian Tunggal, Melaka, Malaysia Article Info ABSTRACT Article history: Received Nov 29, 2021 Revised Jul 2, 2022 Accepted Jul 28, 2022 Cholesterol is a type of lipid found in the human body and is susceptible to abnormalities. It can be detected via lipid profiling through blood sampling. In addition, cholesterol can be detected through the presence of a "sodium ring" in the eye iris called the corneal arcus (CA), presenting a new preliminary detection method that is less invasive. Therefore, this paper proposed a non-invasive method in detecting cholesterol based on convolutional neural network (CNN) model representation using 300 normal and 300 abnormal iris images from UBIRIS and medical web images. In this work, contrast-limited adaptive histogram (CLAHE) and unsharp masking process was applied first on CA images to enhance the quality of CA images. To detect the CA images, the dataset was trained and tested using three pre-trained CNN architectures; one is created from scratch, another are Resnet-50 and VGG-19 architectures that were fine-tuned to the CA images. The best result was exhibited by proposed pre-trained CNN model created from scratch with 10-fold cross-validation that produced high average detection accuracy at 98.81%. Thus, deeper network implementation is recommended in the future to further improve CA localization for optometrists used in their daily clinical tasks in detecting cholesterol. Keywords: Cholesterol Corneal arcus Convolutional neural network Deep learning Non-invasive This is an open access article under the CC BY-SA license. Corresponding Author: Ridza Azri Bin Ramlee Faculty of Electronics and Computer Engineering, Universiti Teknikal Malaysia Melaka 76100 Durian Tunggal, Malaysia Email: ridza@utem.edu.my 1. INTRODUCTION A balanced lifestyle reflects an individual’s physical and emotional well -being, which are inextricably linked in many cases. One of the most common physical health problems is high and low cholesterol in the blood vessels. Excessive cholesterol levels can lead to cardiovascular diseases such as stroke, high blood pressure, and myocardial infarction, hence the need to maintain a healthy blood cholesterol level. An individual’s cholesterol level is determined using a lipid panel or a lipid profile via a bloo d test performed after fasting for nine to 12 hours. However, it is considered an invasive method since its sole purpose, in this case, is to obtain a lipid profile. There are many examples of non-invasive techniques in disease diagnosis. For instance, the x-ray is used to view the images of bones and other structures in the body, while magnetic resonance imaging (MRI) creates accurate images of the body's organs and tissues, and computerised tomography (CT) scan detects abnormalities in the body. Thus, it is highly possible to develop an alternative, non-invasive method for detecting cholesterol for preliminary diagnosis rather than using the lipid profile, even though it does not indicate the level of high-density lipoprotein (HDL) or low-density lipoprotein (LDL) at this stage. With this invention, monitoring cholesterol levels would become more