Localization and diagnosis framework for pediatric cataracts based on slit-lamp images using deep features of a convolutional neural network

Abstract

Slit-lamp images play an essential role for diagnosis of pediatric cataracts. We present a computer vision-based framework for the automatic localization and diagnosis of slit-lamp images by identifying the lens region of interest (ROI) and employing a deep learning convolutional neural network (CNN). First, three grading degrees for slit-lamp images are proposed in conjunction with three leading ophthalmologists. The lens ROI is located in an automated manner in the original image using two successive applications of Candy detection and the Hough transform, which are cropped, resized to a fixed size and used to form pediatric cataract datasets. These datasets are fed into the CNN to extract high-level features and implement automatic classification and grading. To demonstrate the performance and effectiveness of the deep features extracted in the CNN, we investigate the features combined with support vector machine (SVM) and softmax classifier and compare these with the traditional representative methods. The qualitative and quantitative experimental results demonstrate that our proposed method offers exceptional mean accuracy, sensitivity and specificity: classification (97.07%, 97.28%, and 96.83%) and a three-degree grading area (89.02%, 86.63%, and 90.75%), density (92.68%, 91.05%, and 93.94%) and location (89.28%, 82.70%, and 93.08%). Finally, we developed and deployed a potential automatic diagnostic software for ophthalmologists and patients in clinical applications to implement the validated model.

Automatic diagnosis of imbalanced ophthalmic images using a cost-sensitive deep convolutional neural network

“…To evaluate the performance and feasibility of the CS-ResCNN model in detail, we employed four representative feature extraction methods [27, 29] (LBP, WT, SIFT, and COTE), two excellent classifiers [support vector machine (SVM) and random forest (RF)] and three data-level methods [18, 19, 22] [the synthetic minority oversampling technique (SMOTE), borderline-SMOTE (BSMOTE) and under-sampling (UNDER)] for comparison. To achieve the optimal performance of the conventional methods, we firstly presented detailed parameters for classifiers, feature extraction methods and data-level methods as shown in Table 2.…”

“…To achieve the optimal performance of the conventional methods, we firstly presented detailed parameters for classifiers, feature extraction methods and data-level methods as shown in Table 2. Specifically, we chose the parameters of the feature extraction methods and classifiers based on our previous research [27–29]. For the data-level methods (SMOTE, borderline-SMOTE and UNDER), we mainly referred to the previous studies [18, 19, 22] and their open source codes.…”

Current state and future prospects of artificial intelligence in ophthalmology: a review

“…63 Another looked at contact lenses for non-invasively detecting Staphylococcus aureus. 64 Other studies have examined at ophthalmological conditions including cataracts [65][66][67][68] and retinopathy of prematurity 69 whilst some research is focused on improving AI technology for future use, for example, improving AI recognition of the optic nerve head, 70 or improving the false positive or negative rates from imbalanced datasets. 71 Interestingly, some studies are using the ophthalmic datasets for utilization across other aspects of health.…”

Tournament Based Ranking CNN for the Cataract grading

“…After Deep Learning showed impressive performance in various fields, several approach used neural network to grade cataract. Xiyang Liu et al proposed cataract diagnosis framework using convolutional neural network with SVM classifier while Localizing ROI(region of Interest) using Hough transform algorithm [5]. Xinting Gao et al proposed method of Deep learning to obtain high level feature then using svm regressor to grade cataract [6].…”

“…Xinting Gao et al proposed method of Deep learning to obtain high level feature then using svm regressor to grade cataract [6]. Yang et al used ensemble learning for grading cataract with fundus images [7] But these approaches [1,2,3,4,5,6,7]didn't consider characteristic of cataract and medical dataset which can cause degradation of performance. While it is easy to acquire normal case data, getting data of severe illness case is very hard because number of patients are relatively small compared to those who don't have diseases.…”