Python code to detect human facial emotion with convolution neural network(CNN) a deep learning technique using haar cascade classifier. Model folder already consist pre-trained model with 94.7% accuracy to achieve that much accuracy, you need to train the model 100 time or more. You can use testing code with this trained model by giving only the path of the model. The model was trained on the FER2013 data set, which is one of the most challenging datasets for emotion detection.