Implementation of image segmentation networks using pytorch.
Current implemented models: U-Net and Fully-Convolutional DenseNet. See unet and tiramisu folders for details.
Used dataset is from the 2018 Data Science Bowl: Find the nuclei in divergent images to advance medical discovery.
Download the dataset at the following link: https://www.kaggle.com/c/data-science-bowl-2018/data or https://data.broadinstitute.org/bbbc/BBBC038/ .
(tbd bash script for download and unzip)