This project implements a facial emotion recognition system using a pre-trained MobileNet model. The application detects facial expressions in real-time via webcam and classifies them into one of eight categories: Angry, Contempt, Disgust, Fear, Happy, Sad, Surprise, or Neutral.

1. Create a virtual environment:

   conda create -n facial python=3.10 -y
   conda activate facial

2. Install dependencies:

   pip install tensorflow opencv-python

3. Run the application:

   python main.py

This script runs the real-time facial emotion recognition application. It uses OpenCV to capture video frames from the webcam and detects faces using a Haar cascade classifier. Detected faces are preprocessed and passed to the MobileNet model for emotion prediction.

This Jupyter Notebook contains the training and evaluation code for the MobileNet-based emotion recognition model. It details data preprocessing, model training, and saving the trained model as 20_epochs_mobilenet_reloaded.keras.

• The pre-trained MobileNet model is saved as 20_epochs_mobilenet_reloaded.keras.
• Ensure the model file is placed in the same directory as main.py or update the file path in the script accordingly.

• Press ctrl + q to quit the application during runtime.
• Ensure your webcam is connected and accessible.

• Python 3.10
• TensorFlow
• OpenCV

• No webcam feed:

  • Ensure your webcam is properly connected and accessible.

• Missing model file:

  • Verify that 20_epochs_mobilenet_reloaded.keras exists in the correct directory.

• Errors related to OpenCV or TensorFlow:

  • Reinstall the dependencies using the command: pip install tensorflow opencv-python.