READ ME.txt

Requirements:
- Python 3.9.7
- Libraries and their version used during development:
	- tensorflow == 2.7.0
	- pandas == 1.3.4
	- numpy == 1.21.4
	- matplotlib == 3.5.1
	- scikit_learn == 1.0.2
	- scipy == 1.7.2
	- opencv_python == 4.5.4.58
- Dataset:
	- fer2013 (**included in this folder**)
	- link: https://www.kaggle.com/ashishpatel26/facial-expression-recognitionferchallenge
- CNN Model and Weights:
	- model.json (**included in this folder**)
	- model.h5 (**included in this folder**)
- Cascade Classifier: 
	- haarcascade_frontalface_default.xml (**included in this folder**)
	- link: https://github.com/opencv/opencv/blob/master/data/haarcascades/haarcascade_frontalcatface.xml

Instructions:
- Model training and validation is done in "FED_Train_Validation" notebook file
	- It includes data preparation, image augmentation, model building, training,
	  testing/validation showing loss and accuracy of both, and confusion matrix for
	  model evaluation.
- The model is implemented in "FED_app_run" notebook file using OpenCV library that captures
  live images from device camera and predicts the emotion and displays with accuracy.
- The application can be only closed using 'q' key in the keyboard.

** As model weights file and dataset are missing you can download from: https://drive.google.com/file/d/1olyGHehyu5rtf0OamtqP8AaO4LiHo-Ac/view?usp=sharing
   or download the dataset from the link above and train the model and save weights file as "model.h5" **