CataractCLSNet: Adaptive GAN-Based Image Augmentation for Cataract Segmentation and Classification Model using Deep Learning with Attention Mechanism

One of the most common vision impairments around the world is cataracts. However, artificial intelligence in the research field of cataracts is still developing, especially when compared to other important age-related eye diseases like Age-related Macular Degeneration (AMD), glaucoma, and diabetic retinopathy. However, a number of additional research teams suggested a novel artificial intelligence-based cataract assessment for intraocular lens power prior to cataract extraction. Although early encouraging performances were noted by the majority of drawbacks, including a lack of independent validations, robust, and superior training data. So, in this research work, an effectual cataract detection model using a supervised learning strategy is developed to overcome such challenges. Initially, the images needed to perform the cataract classification are taken from benchmark sources. Then, the acquired images are carried to the image augmentation process using the Adaptive Generative Adversarial Network (AGAN), where the parameters are tuned by the Modified Dung beetle optimizer (MDBO) to improve the augmentation performance. Once the image augmentation process is completed, the augmented images are used in the segmentation process using the Dilated Mask Region-based Convolutional Neural Networks with Attention Mechanism (DMR-CNN-AM). Then, the outcome obtained through the developed DMR-CNN-AM model is fed into the EfficientnetB7 with Gated Recurrent Unit Layer (ENB7-GRU)-based cataract classification model. Finally, the ENB7-GRU model provided the classified output, and the outcome is compared with the Synthetic Minority Over-sampling Technique (SMOTE) model to prove the superior outcome of the suggested model compared to SMOTE. The performance of the suggested model is contrasted with various baseline models to confirm its effectiveness. The World Health Organization (WHO) and other organizations conduct comprehensive surveys to estimate the global burden of visual impairment, including cataracts. Countries like India often conduct national eye health surveys to assess the burden of eye diseases, including cataracts, and plan effective interventions for prevention and treatment.