Nystagmus Patterns from Eye-Tracker Data
Zarib Müller · 2024
By analyzing eye fixations, we gauge the difficulty of interpreting fixated information and the percentage of fixation time per question, shedding light on cognitive processing. Shifting gears, the analysis of Nystagmus waveforms from eye-tracking records plays a pivotal role in clinically interpreting this pathological movement. However, automating this analysis faces challenges due to the presence of natural eye movements and eye blink artifacts, which intertwine with the signal of interest. To address this, we propose a Convolutional Dictionary Learning-based method that automatically highlights Nystagmus waveforms, effectively separating natural motion from pathological movements. Our approach demonstrates improved pattern recovery rates, supported by clinical examples illustrating its efficacy.