soobinrho / emotion-recognition-website

This is a demo website for COSC380 Artificial Intelligence.

How to Run

# Install pnpm.
curl -o- https://raw.githubusercontent.com/nvm-sh/nvm/v0.39.5/install.sh | bash
nvm install node
npm install -g pnpm
pnpm setup

# Clone this repository.
git clone https://github.com/soobinrho/emotion-recognition-website.git

# Install dependencies.
cd emotion-recognition-website
pnpm install

# Run a development server.
pnpm dev

Development Logs

Warning

Work in progress. Model integration underway.

• Trained the model on Jupyter Lab.

model.save("final_model.keras")

• Converted the model into TF.js layers format. This required installation of a pyton tensorflowjs package. Do not install this in the same environment as the model's because conflicts can occur, which tend to break the entire training process.

conda deactivate
conda create --name tfjs python=3.10
conda activate tfjs
pip install tensorflow[and-cuda] Cython tensorflowjs

# Source:
#   https://www.tensorflow.org/js/tutorials/conversion/import_keras#step_1_convert_an_existing_keras_model_to_tfjs_layers_format
tensorflowjs_converter ./my_final_model ./ml_model

• Used react-webcam package for the webcam functionality.