A Positive/Unlabeled Approach for the Segmentation of Medical Sequences using Point-Wise Supervision

Description --- This archive contains data and results (model checkpoint, segmentation results, ...) used and obtained in paper: Laurent Lejeune, Raphael Sznitman, A Positive/Unlabeled Approach for the Segmentation of Medical Sequences using Point-Wise Supervision, Medical Image Analysis, 2021 (in preparation) Details --- Each directory is named DatasetXY, where X denotes the type (surgical instrument, cochlea, slitlamp, and brain), and Y denotes the index of the sequences (from 0 to 3). For the data.zip archive, we provide a CSV file that contains the 2d locations used in our experiments. For the results.zip archive, the checkpoint of our models are in PyTorch format.