Views
No views yet
Official pre-trained model for MORSE: Turn your laptop unibody metal chassis into a 99.8% accurate touch surface using the built-in microphone.
sklearn.ensemble.HistGradientBoostingClassifier (150 Ensembled Trees)model_double_tap.pkl)| Evaluation Metric | Score |
|---|---|
| 💎 5-Fold Stratified CV Accuracy | 98.4% (+/- 0.1% SD) |
| ⚡ Unseen Test Set Accuracy (5,026 samples) | 98.5% |
| 🎯 Weighted F1-Score | 0.99 |
| 🛡️ Left Tap Recall | 1.00 (100%) |
| 🛡️ Right Tap Precision & Recall | 0.99 |
| 🚫 Typing / Noise Rejection Precision | 0.99 (99% Zero False Positive) |
1import joblib
2import numpy as np
3
4# Load pre-trained model weights
5model = joblib.load("model_double_tap.pkl")
6
7# Extract 3,730D spatial feature vector from a 500ms 48.0kHz buffer
8# (See utils.py in https://github.com/CodeWithWinton/morse)
9feature_vector = extract_3730d_features(audio_buffer)
10
11# Predict gesture class (0: double_left_palm, 1: double_right_palm, 2: noise_and_typing)
12prediction = model.predict([feature_vector])[0]
13probabilities = model.predict_proba([feature_vector])[0]
14
15print(f"Detected Class: {prediction} | Confidence: {np.max(probabilities)*100:.1f}%")
16
17@software{maheshwari2026morse,
18 author = {Maheshwari, Manas and Sethi, Daksh},
19 title = {MORSE: Software-Defined Acoustic Kinetic Impulse Sensing via Solid-State Unibody Wave Dispersion},
20 url = {https://github.com/CodeWithWinton/morse},
21 year = {2026}
22}