OWLET
Validation and technical details
![](images/owlet_demo.png)
OWLET combines algorithms from computer vision, machine learning, and ecological psychology to estimate infant's point-of-gaze. I validated OWLET using data from a large sample of 6-8 month-old infants. A few key findings were:
- The average x/y spatial accuracy was 3.36°/2.67° across infants.
- There were no differences in data quality for webcams vs smartphones.
- We observed greater racial/ethnic and diversity when families were given the option to participate in remote studies using their smartphones.
![](images/owlet1.png.webp)
![](images/owlet2.png.webp)
![](images/owlet3.png.webp)
All modules, source code, and instructions for using OWLET can be found on GitHub.
Mac OSx App
I developed a beta version of a MacOSx App to better support users who are unfamiliar with using Python virtual environments. To request a password to download the app, please fill out the form below.