Geometric Gesture Classifier
Real-Time Single Hand Gesture Recognition
RUN
This is a real-time hand gesture recognition tool, running entirely in your browser with no data leaving your device. It uses MediaPipe Hands to extract 21 3D landmarks per hand, then classifies the gesture from those landmarks with a simple, fast, rule-based classifier — no training or model download involved, it just checks finger geometry directly (which fingers are extended, and how far apart the thumb and index fingertip are) against a fixed set of rules for the 8 gestures below.
The following 8 gestures are recognised out of the box:
- Fist ✊
- Open Hand ✋
- Point ☝️
- Peace ✌️
- Thumbs Up 👍
- Rock 🤘
- Call Me 🤙
- OK 👌
The data is exported as a CSV file structured as follows:
| frame | timestamp | tag | hand | gesture | confidence | joint_0_x | joint_0_y | joint_0_z | … | joint_20_x | joint_20_y | joint_20_z |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 42 | 1718000000 | baseline | Right | Peace | 0.913 | 0.512 | 0.748 | -0.031 | … | 0.421 | 0.302 | -0.089 |
Landmark coordinates are in MediaPipe's normalised image space (x, y ∈ [0, 1]). The z value is depth relative to the wrist (negative = closer to camera). Rows where no hand is detected are omitted.
Reference: Google. MediaPipe Hand Landmarker. developers.google.com/mediapipe/solutions/vision/hand_landmarker