Implementation of a multimodal brain tumor segmentation technique based on 3D Unet encoder-decoder architecture A variational autoencoder-based regression model to predict the days of survival with a low error rate Part of the project done for the Advanced Machine Learning (EECE 7397) course at Northeastern University taught by Prof. Jennifer Dy.