Your Eye Tracker is a low-budget eye tracking platform based on a cheap endoscope camera,
3d printed parts and OpenCV image processing. It currently reaches around 2.5° error on average, with a minimum of 5°.
This is a little less accuracy as than what commercial eye trackers offer, at
YET can be assembled from scratch within minutes. Provided one first prints the socket. The tutorial describes the assembly, installation and usage of YET. Yet comes ready to try, with a pre-defined experiment. Read more on the Uncanny Valley experiment[Use download button]
Built-in Data Science
Yet Zero uses CSV tables to specify the experimental design, which can be edited using any table calculation program that exports to CSV, or any text editor. Included is a full data analysis pipeline in tidy R, using the Quarto system. Beginners can quickly produce research reports and learn the basics of eye tracking data analysis with Yet.
Theory of Operation
The inspiration for Yet came from an evolutionary point of view, more specifically the cooperative eye hypothesis. Basically, the theory says that the human eye evolved into its modern appearance to make it easier for others to see where we are looking. If the eyeball is such a strong signal, then building an eye tracking mechanism should be easy, right? If you are interested, here is a deeper discussion of the theory[Use download button]