HandyTrak: Recognizing the Holding Hand on a Commodity Smartphone from Body Silhouette Images
summary
HandyTrak is a software pipeline that takes front-facing camera silhouettes, segments body outlines, and feeds them to a user-dependent CNN to classify left, right, or both-hand grips. It runs at 2 Hz on commodity phones, requires ~3 minutes per-user training, and achieves about 89.03% average accuracy (SD 8.98%) across five camera positions, three usage scenarios, and two grip poses. The system improves with longer windows and trades continuous camera use for battery and privacy costs.
AI-written summary · about
chapters (6)
- 0:00Intro
- 0:38Motivation
- 2:02Related Work Automatic Handedness Detection
- 4:52System Design System Overview of Handy Trak
- 5:42Device Setup Different Camera Configurations
- 7:02Discussion
related talks
07mFingerTrak: Continuous 3D Hand Pose Tracking by Deep Learning Hand Silhouettes Captured by ...
19mBackHand: Sensing Hand Gestures via Back of the Hand
20mToward Realistic Hands Gesture Interface: Keeping it Simple for Developers and Machines
06mWR-Hand: Wearable Armband can Track User's Hand
16mInvestigating the feasibility of finger identification on capacitive touchscreens using deep ...
14mFinger-Aware Shortcuts
14mDeepFisheye: Near-Surface Multi-Finger Tracking Technology Using Fisheye Camera
14mBack-Hand-Pose: 3D Hand Pose Estimation for a Wrist-worn Camera via Dorsum Deformation Network