Skip to content
SDGs Center UDINUS

Evidence record · UDINUS · SDGs

YOLOVX Method for Cataract Early Detection

Rektorat UDINUS 2025 Published 2026-10-30

Evidence record · UDINUS · SDGs

Unit
Rektorat UDINUS
Period
2025
Goals
Record
e5301811-2db9-4174-a862-29bdafd44075
Published
2026-10-30

Universitas Dian Nuswantoro conducted an activity focused on developing a deep learning-based model for early cataract detection. The activity used YOLOv10x, an object detection algorithm described in the source as known for its speed and accuracy. The activity participants were Widyananda, Muhammad Asdar; Paramita, Cinantya; Supriyanto, Catur; Wibowo, Angga Wahyu; Utomo, Danang Wahyu; and Widyaatmadja, Swanny Trikajanti.

The activity addressed cataract, which the source identifies as one of the leading causes of blindness worldwide, particularly in developing countries with limited access to medical services. The dataset consisted of images of eyes showing various cataract conditions. These images were obtained from open sources such as Kaggle and Roboflow Universe. Before model training, the data underwent normalization, augmentation, and annotation. The source states that these processes were used to ensure high quality in model training. The model was trained using high-performance hardware.

The evaluation used mAP, F1-Score, Precision, and Recall as metrics. The experimental results reported that the model achieved its highest accuracy of 96.8% at the 100th Epoch. The source presents YOLOv10x as a superior solution for early cataract detection. It also states that applying the model can enable cataract detection to be performed quickly, accurately, and affordably through devices such as mobile phones. According to the source, this application contributes to blindness prevention in resource-constrained areas. The research is further described as an innovative approach to reducing the public health burden and improving the quality of life of affected individuals.

In relation to SDG 3, the UDINUS activity represents health research addressing cataract detection and blindness prevention through a technology-based approach. Its scope includes dataset preparation, model training, and evaluation of an object detection model for early detection.

Reference https://www.scopus.com/pages/publications/105012761285?origin=resultslist