Gesture Controller

A deep learning powered multimedia controller operated by hand gestures.

The application is built using Flask, Python, OpenCV, Mediapipe, and TailwindCSS, with a TensorFlow-trained model and data collected via Mediapipe.

About

The model consists of two dense layers with ReLU activation, followed by a fully-connected dense layer with softmax activation. It uses the Adam optimizer and sparse categorical cross-entropy as the loss function. The model achieves a validation accuracy of 97%.

Landmark data was collected from the HaGRID (512px) dataset.

Setup

1. Clone the repository:

   git clone https://github.com/siddhp1/Gesture-Controller.git
   cd Gesture-Controller/app

2. Create environment and install dependencies:

   python -m venv venv
   source venv/bin/activate
   pip install -r requirements.txt

Usage

1. Run application:

   python -m main

2. Open GUI:

   Go to http://localhost:5000 in your web browser.

License

This project is licensed under the MIT License.