Deep Rock Paper Scissors Gesture Recognition

In this repository you will find scripts to:

• Recognize hands using Google MediaPipe.
• Record a gesture dataset.
• Train a gesture classifier using Tensorflow and Keras.
• Implement a Rock Paper Scissors recognizer using Google MediaPipe and the trained model.

Prerequisites

• Python 3.8.x or newer
• Python venv

Preparation

Fetch required resources

Create a virtual enviroment:

python3 -m venv env
source env/bin/activate
pip3 install -r requirements.txt

Rock Paper Scissors Gesture Recognition

With the venv opened, run:

python3 rock-paper-scissors.py

Train your own model

Record your dataset

With the venv opened, run:

python3 record-dataset.py

Train the gesture classifier

With the venv opened, run:

python3 train-gesture-classifier.py