Stanford University's Center for Artificial Intelligence in Medicine and Imaging (AIMI) has teamed up with Microsoft to launch a new platform that will host and share imaging datasets for the training of artificial intelligence (AI) algorithms.
In addition to creating an open and global repository of datasets from Stanford and other institutions around the world, the platform will also provide a hub for sharing research and will even offer cloud-based computing power for researchers, according to a post by the university.
The shared dataset platform can be accessed here. The collection of deidentified and annotated datasets are available in three categories: bone, extremity, and chest.
The platform already has nine datasets containing more than one million images, and AIMI co-director Dr. Matthew Lungren predicts that will more than double over the next year. In the future, the university said it plans to also offer standardized machine-learning tools and pretrained models leveraging open-source data and common architectures.
