EclipseTouch: Touch Segmentation on Ad Hoc Surfaces using Worn Infrared Shadow Casting
summary
Headset-mounted IR illuminators plus an IR camera produce structured shadows that encode finger-to-surface distance; a neural model trained on per-LED shadow crops classifies hover and contact and regresses distance. Prototype uses up to four LEDs, a 400 FPS camera (five-frame sequence at 80 FPS), and leave-one-person-out evaluation across 12 materials and lighting conditions, yielding 6.9 mm mean hover error and 98.0% touch accuracy with no calibration and real-time inference.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 1:12Motivation for touch input
- 3:58Live demonstration
- 5:49How Eclipse Touch works
- 6:53System implementation
- 8:15Data collection and evaluation
- 9:57Results and conclusion
- 10:18Q&A session
related talks
06mShadowSense: Detecting Human Touch in a Social Robot Using Shadow Image Classification
18mInfiniTouch: Finger-Aware Interaction on Fully Touch Sensitive Smartphones
14mDeepFisheye: Near-Surface Multi-Finger Tracking Technology Using Fisheye Camera
19mSymmetriSense: Enabling Near-Surface Interactivity on Glossy Surfaces using a Single Commodity ...
05mReady, Steady, Touch! — Sensing Physical Contact with a Finger-Mounted IMU
17mAccurate and Low-Latency Sensing of Touch Contact on Any Surface with Finger-Worn IMU Sensor
17mPredicTouch: A System to Reduce Touchscreen Latency using Neural Networks and Inertial ...
20mSkinTrack: Using the Body as an Electrical Waveguide for Continuous Finger Tracking on the Skin