Emotion Detection Model using Deep Learning Techniques

• Detecting and Classifying emotions from human speech samples.
• Used Librosa package in python to load and pre-process audio data, performed feature engineering.
• Explored different Deep Learning Convolutional Neural Network models using ML libraries Keras and TensorFlow.
• Built a final model using a modified VGG-16 network Architecture with an accuracy of 65% in predicting 8 different emotions as compared to a human accuracy of 60%.

Files:
Emotional_Intelligence_Model.ipynb - Jupyter notebook that contains the code for importing, processing data and training the Neural Network.
Emotional_Intelligence_Testing.ipynb - Jupyter notebook for testing the model on new data. Emotion_Detection_DeepLearning.pdf - Project Report.
Audio.ipynb - Jupyter notebook for capturing audio samples.