Pyro: Thumb-Tip Gesture Recognition Using Pyroelectric Infrared Sensing
summary
Passive pyroelectric infrared sensing recognizes six thumb-tip gestures by capturing thermal signals with a Fresnel-lensed PIR, analog front-end, and a microcontroller streaming to a laptop for feature extraction and random-forest classification. A 10-participant dataset (3,600 samples) yields 93.9% cross-validation and 84.9% leave-one-session-out accuracy. Prototype fits on a watch and glass-mounted unit, senses from 0.5–30 cm, handles varied lighting and some background motion, and fails when finger temperature equals room temperature.
AI-written summary · about
chapters (12)
- 0:00Introduction to Pyro
- 1:12Motivation for Pyro
- 2:17Pyroelectric infrared sensing benefits
- 3:45How PIR sensors work
- 4:52Sensing principle of Pyro
- 5:55Demo of gesture recognition
- 7:07Pyro system implementation
- 8:10Accuracy evaluation results
- 9:15Robustness against noise
- 12:25Demo applications
- 13:54Limitations and future work
- 14:34Take-home messages and Q&A