Thesis Project Form

Moderate strabismus can disrupt binocular correspondence and promote suppression of the weaker eye, particularly when accompanied by amblyopia. This thesis tests whether relocating the region of highest effective, task-relevant image quality can bias vergence and ocular alignment, creating a balanced virtual match between the dominant fovea and an artificially privileged eccentric locus in the non-dominant eye.

The project will develop an AR-based dichoptic system that applies real-time, gaze-contingent, space-variant image transformations. In the non-dominant eye, the locus of maximal effective image quality will be shifted away from the fovea; the dominant-eye foveal image will be degraded to match the quality available at the selected eccentric locus and reduce suppression. Activities include implementation in Unity/C#, binocular calibration, measurement of eye-tracking accuracy and end-to-end latency, and design of psychophysical experiments. The main hypothesis is that artificial-fovea offsets induce compensatory eye rotations in the opposite direction. Ocular alignment and vergence will be analysed as functions of offset direction, magnitude and image-quality balance, with conventional contrast-balanced dichoptic stimulation as a control.

The student will gain practical experience in Unity/C#, AR/VR rendering, eye tracking, gaze-contingent stimulation, real-time image processing and binocular psychophysics. Training will cover headset and oculomotor calibration, latency and spatial-accuracy assessment, experimental design, analysis of vergence and eye-position data, statistical modelling and responsible human-participant testing.