Multiple Major Ocular Disease Classification Using Deep Learning Methods: A Comprehensive Review

Ocular diseases have become an important issue worldwide. Every year, various ocular diseases affect numerous individuals, leading to blindness for many. Despite seeking treatment in hospitals or from doctors, accurately diagnosing the disease remains a challenge. In numerous cases, the correct diagnosis is elusive and time-consuming, leading to unnecessary delays. This is why computer-aided diagnosis is necessary to swiftly diagnose the disease. However, most Computer-Aided Diagnoses (CADs) are designed for single diagnoses. Therefore, there is a crucial need to develop CAD technology for detecting multiple ocular diseases. Many researchers are currently investigating this issue and developing various deep-learning models to find a solution. They have developed many techniques and models through which various types of ocular diseases can be diagnosed quickly and accurately. So far, the highest AUC score achieved is 99%. Reading all the literature on these techniques is time-consuming, so a literature review paper on this topic is necessary. In this review paper, the techniques from some important papers on this topic have been summarized. A critical analysis has been provided of all the papers, including their limitations and potential solutions.