FACIAL EXPRESSION RECOGNITION SYSTEM USING DEEP LEARNING (Machine Learning)
--> Machine Learning program is developed using deep convolution neural network. We have used theano,PIL
and scipy libraries of Python for implementation of this ML program where seven different expressions of face are
recognized based on trained data set.


Libraries Used :-
1.Theano.
2.PIL
3.Numpy
4.cPickle
5.Gzip
6.sklearn


Here two models are there-:
1. Convolution layer(deep learning)
2. SVM



Dataset Used- jaffer(213 images) -converted into pickle format using Converter.py and processed using crop.py
net3.py has been used in the deep learning algorithm for buiding the desired convolution layer and which uses the concept of feature mapaping
Here codeFACIAL is the main code which builds the main trained model of deep learning using 6 convolution layer and 2 fully connected layers
SVM model is build by loading the same jaffe data which has been built using Converter.py in pickle format and then fitting the training
model and label to predict the accuracy of the code .
Accuracy is better if size of the dataset is large and it has been preprocessed by any feature extraction model on the dataset .
We had not applied the feature extraction on the code. If you are familiar with the same, you can append the logic of same in this model.