A-Eye Tracker: Human Eye Defect Tracker and Analyzing Software

Eye tracking is a valuable method for detecting and analyzing eye movement in various research fields and healthcare environments. It can assist in patient diagnosis and treatment. However, the accuracy and reliability of eye-tracking tools have been a long-standing challenge in the field. This project presents a software application that utilizes DeepLabCut (DLC), a deep learning-based computer vision library, to track eye movements from a live stream video or a prerecorded video (.mp4) file. The application gives clinicians and researchers an affordable, reliable, and portable tool for eye tracking. This helps them identify and analyze eye defects more effectively. This software can generate useful outputs like eye movements as CSV data, and videos showing different labelled parts of the eye, and help doctors decide the most suitable treatment options over time. Additionally, this project showcases an example of a software application & its features built using the Dash Python framework. It is integrated with DLC API, which runs as a web service. Together they create an effective tool for eye tracking. This application offers an option to integrate a Convolution Neural Network (CNN) model into the web app resulting in customized features. It helps researchers to utilize it in their respective fields of study. The practical finding indicates that the developed application is highly effective in providing accurate and real-time eye-tracking capabilities. In summary, this project explores the difficulties and possible solutions in eye tracking using DLC. It also offers valuable insights into the future to research in this field. This project ‘A-Eye Tracker’ (Phonetically similar to AI).