Paper Contents
Abstract
Cataract is the most common cause of preventable blindness especially in rural and underserved areas where there is no access to sophisticated diagnostic equipment. The proposed research is set to create a low-cost and transportable cataract detection device based on eye images taken with smartphones. To enhance the quality of images, preprocessing methods, such as denoising, resizing, and contrast enhancement, were used. On Google Colab, three deep learning models MobileNetV2, InceptionV3, and EfficientNetB0 were trained and tested. Findings revealed that MobileNetV2 was the fastest with an accuracy of 94%, InceptionV3 came in second with 91 and EfficientNetB0 had the lowest with 50. These results show that lightweight architectures such as MobileNetV2 are portable to mobile and provide a scalable, cost-effective early cataract screening solution in low-resource environments.
Copyright
Copyright © 2025 Ms. Pooja Kadam. This is an open access article distributed under the Creative Commons Attribution License.
