The Problem The noise level in a modern video-based eye-tracker is usually larger than the size of the smallest saccades. Thus, it can be challenging to discriminate small eye movements… Click to show full abstract
The Problem The noise level in a modern video-based eye-tracker is usually larger than the size of the smallest saccades. Thus, it can be challenging to discriminate small eye movements from system noise. Various methods have been proposed, but most require hand-tuning of parameters to achieve good performance. We would like a method that automatically adapts to the statistics of the input signal.
               
Click one of the above tabs to view related content.