CYTOMINE : a web platform for collaborative analysis of multi-gigapixel images with machine learning.
Cytomine (http://cytomine.be, source code on https://github.com/cytomine) is, to the best of our knowledge, the first open-source rich internet application to enable highly collaborative and multidisciplinary analysis of multi-gigapixel imaging data.
Built at the University of Liège with state-of-the-art RESTful (Grails), noSQL (MongoDB, Redis) and automated deployment (Docker) technologies, it supports both remote visualization, collaborative semantic annotation, and semi-automated analysis through the web, making it an ideal tool for collaborative research, teaching and diagnosis in every large-image related topics.
Its design was driven by life science and bioimage informatics research needs: software versatility, interoperability, modularity and extensibility, image recognition tailored via learning from ground-truth data and proofreading tools, reproducible research, and data accessibility and reusability.
It is being used for years by our collaborators working with large sets of bioimages in numerous domains including cancer research, development, and toxicology, and is now adapted for pedagogical purposes.
Overall, we believe Cytomine is an important new tool of broad interest to foster active communication and distributed collaboration between life, computer and citizen scientists, but also physicians, teachers and students, in an unprecedented way using machine learning and web communication mechanisms.
After 5 years of active developpement, use in research and education, our team will be proud to annonce at FOSDEM that the entire source code will be available at this date under an Apache (v2) licence.
Our lightning talk will focus on Cytomine main features, illustrated by research and teaching practical examples, and will broach how we want to carry on our project in the future.