This paper presents an automated detection of pterygium disease in anterior segmented photographed images. The five main steps are implemented as follows; pre-processing of the ASPI database, model development using a proposed deep convolutional neural network (DCNN) architecture, evaluation of the tested models, analysis of the models' efficiencies and comparison of model performance. The proposed DCNN architecture consists of six learned layers that are three convolutional and three fully connected layers. This experimental work focuses on the identification of the best combination of hyperparameters for the proposed DCNN architecture. The optimal hyperparameter combination with learning rates of 0.001 and 0.0001 and epoch values of 20 and 25, provides the best performance of 94.09% accuracy, 93.93% F1 score, and 88.32% Matthew correlation coefficient with the highest area under the curve value for an average fold of 95%. The results prove that the proposed DCNN architecture has a promising potential to be applied in the development of future pterygium screening tools.

No abstract

Keratoconus (KC) is a condition of the bulging of the eye cornea. It is a common non-inflammatory ocular disorder that affects mostly the younger populace below the age of 30. The eye cornea bulges because of the conical displacement of either outwards or downwards. Such condition can greatly reduce one’s visual ability. Therefore, in this paper, we afford a mobile solution to mitigate the KC disorder using the state-of-the-art deep transfer learning method. We intend to use the pre-trained VGGNet-16 model and a conventional convolutional neural network to detect KC automatically. The experimental work uses a total of 4000 side view lateral segment photographed images (LSPIs) comprising 2000 of KC and non-KC or healthy each involving 125 subjects. The LSPIs were extracted from the video data captured using a smartphone. Fine tuning of three hyperparameters namely the learning rate (LR), batch size (BS) and epoch number (EN) were carried out during the training phase to generate the best model of which, the VGGNet-16 model fulfilled it. For the KC detection task, our proposed model achieves an accuracy of 95.75%, a sensitivity of 92.25%, and specificity of 99.25% using the LR, BS and EN of 0.0001, 16, and 70, respectively. These results confirmed the high potential of our proposed solution to apprehend KC prevalence towards an automated KC screening procedure.