Automatic visual impairment detection system for age-related eye diseases through gaze analysis

Area of Science:

• Ophthalmology
• Biomedical Engineering
• Computer Science

Background:

• Age-related Macular Degeneration (AMD) causes central vision loss (scotoma), impacting daily activities.
• Current visual field assessment tools like Amsler grid and Microperimetry are manual and have limitations.
• Objective and automated methods are needed for reliable visual impairment detection.

Purpose of the Study:

• To develop and validate an automated system for detecting visual impairment using gaze tracking.
• To assess the efficacy of the AVIGA system in differentiating visual impairment severity.
• To compare AVIGA's performance against established visual assessment methods.

Main Methods:

• Development of the Automated Visual Impairment Gaze Tracking (AVIGA) system.
• Implementation of Impulse Stimuli Response (ISR) and Pursuit Stimuli Response (PSR) tests within AVIGA.
• Application of Support Vector Regression (SVR) for analyzing assessment results and classifying impairment severity.

Main Results:

• The AVIGA system demonstrated strong correlation with visual acuity (VA) tests.
• AVIGA showed superior performance in identifying the presence of visual impairments compared to Microperimetry.
• The SVR-based approach effectively differentiated the severity of visual impairment.

Conclusions:

• The AVIGA system offers a promising automated approach for visual impairment detection.
• Gaze tracking combined with stimuli response tests provides a reliable method for assessing visual field changes.
• AVIGA has the potential to improve the diagnosis and management of visual impairments, particularly those associated with AMD.