SimplePupil - a LabVIEW-based pupil tracker for neuroscience applications
In recent years there has been much interest in neuroscience in using pupil size as an index of arousal state. Unfortunately, available pupil trackers are not particularly suited for this purpose, leading many labs to create their own pupil tracking software. Our lab followed suit, resulting in SimplePupil. In creating this program we sought to avoid common pitfalls:
1) Rather than use an opaque circle-finding algorithm (such as MATLAB's imfindcircles) that is difficult to troubleshoot, we created a transparent pipeline, which consists of smoothing the image, thresholding it, and setting criteria for what qualifies as a circle. Since each step is visible to the viewer, it is clear which parameters may need to be changed and how. We also found this approach to be considerably faster and more versatile than circle-finding.
2) Rather than only record online and analyze offline, which can produce videos that are later found to be of insufficient quality for analysis, we created a recording tool that comes with the same pipeline. This allows us to quickly optimize physical parameters such as the camera location and lighting before recording.
3) Rather than only move forward through the video in offline analysis, we created a slide bar tool so one can easily navigate backwards and forwards with the pupil marked to make sure analysis works as it should.
If you are interested in obtaining this program and also receive more information on its use and integration into a neural recording setup, please contact firstname.lastname@example.org. This tool was created by Michael Sokoletsky, Akiva Rappaport, and Ori Noked.