Emotion_detection

This projects aim to classify real time facial emotion recoginition to either of one of the universal seven emotions using Deep convolutional neural network(CNN) learning and open cv

Installation using command line using pip

1. pip install opencv-python
2. pip install tensorflow
3. pip install keras

Installation using Anaconda prompt

1. conda install opencv-python
2. conda install tensorflow
3. conda install keras

Note:

If you are using GPU then uncomment this code block in model.py

'''config = tf.compat.v1.ConfigProto()

config.gpu_options.per_process_gpu_memory_fraction = 0.15

session = tf.compat.v1.Session(config=config)'''

To run the project :

1. First extract train folder and test folder from this link and get this to the main project folder.

2. open command line or anaconda prompt from this main project folder and type python main.py

3. open your Google chrome and go to localhost:5000 and run your project

Project Architecture

1. 4 CNN layer with 64,128,256 512 filter respectively
2. 3 dense layer with 256, 512 and 7 nodes respectively
3. model is converted to json and implemented in flask framework