This is an Introduction to AI project that was developed by Robert Sika and John Adenu-Mensah. The essence of this project is to create an unsupervised AI model that is able to predict emotions using a Recurrent Neural Network model Long Short Term Model (LSTM)

• Features
• How To Use
• Demonstration

The emobot uses the Tess dataset to train and test it's accuracy and validation. It used MFCC to extract features from the audio files and those features were used in the training and testing process using an LSTM and some Dense models which used various activation functions like relu and softmax and the rmsprop optimizer. The application has a user-friendly interface which was made using Streamlit, a lightweight python library that can be used to host web applications, The aplication allows a user to upload a file and then predict the emotion that is found in the uploaded file.

To use our web application you can follow these simple steps:

Access our application by typing the following in your terminal

streamlit run deployment.py

and upload a 1 to 3 second long file in the form OAF_test_[emotion] or YAF_test_[emotion]

Click 'Browse Files' and select a file that you want to use to predict.

Click the 'Predict Emotion' button and wait for an emotion to be predicted.

Model is highly effective with audio files that are taken from the official Tess dataset, and is likely to predict disgust for most audio files uploaded.

You can find the link to our implementation and presentation at the links below.

https://youtu.be/oBahaY4CvjI

https://youtu.be/Bkd-BZnkWFY

Thank you!