One of the many conditions that can affect the eyes is called a cataract. Cataracts are an extremely common condition, but unfortunately, they frequently result in the irreversible loss of an eye. A cataract may develop for a number of different reasons. The early identification of cataracts is beneficial to the therapy process. Surgery may also be necessary in cases where the
cataract is caused by a serious infection. Therefore, the patient will benefit from an early diagnosis of the condition. The identification of cataracts will begin with the capture of fundus pictures of both the right eye and the left eye. We dealt with around 1094 datasets, 594 of which contained photos with cataracts, while the other datasets had normal photographs. We used four algorithm models: vgg19, vgg16, Inceptionv3, and Restnet50. Vgg19 achieved an accuracy of 95%, vgg16 achieved an accuracy of 94%, Inceptionv3 achieved an accuracy of 85%, and Restnet50 achieved an accuracy of 87%. The validation accuracy assessed how effectively the models generalized to new data, which was used to avoid overfitting.