Motivation: Collaborative analysis of massive imaging datasets is essential to enable scientific discoveries. Results: We developed Cytomine to foster active and distributed collaboration of multidisciplinary teams for large-scale image-based studies. It uses web development methodologies and machine learning in order to readily organize, explore, share and analyze (semantically and quantitatively) multi-gigapixel imaging data over the internet. We illustrate how it has been used in several biomedical applications. Availability and implementation: Cytomine (http://www.cytomine.be/) is freely available under an open-source license from http://github.com/cytomine/. A documentation wiki (http://doc.cytomine.be) and a demo server (http://demo.cytomine.be) are also available. Contact: info@cytomine.be Supplementary information: Supplementary data are available at Bioinformatics online.

We present a novel methodology combining Web-based software development practices, machine learning, and spatial databases for computer-aided quantification of regions of interest (ROIs) in large-scale imaging data. We describe our main methodological choices, and then illustrate the benefits of the approach (workload reduction, improved precision, scalability, and traceability) on hundreds of whole-slide images of biological tissue slices in cancer research.

Digital slide scanning is advancing the field of pathology and biomedical research, resulting in very large amounts of imaging data. From the computer science point of view, it is challenging to efficiently share, annotate and analyze such data due to their distinct geographical localizations, their high dimensionality, and their numerous sources of variability (scanning equipments, file formats, acquisition protocols, application domains, ...).

We integrated recent web technologies and generic methods to ease visualization and annotation of large-scale images through the Web [1]. On server-side, our data model allows to create and manage projects, where each project corresponds to a specific experimental study. A project is described by a list of authenticated users, a list of images, an user-defined, structured, ontology, and annotations (ROI) drawn by users and described by ontology terms. All data are stored in a spatial, relational, database and accessible through a RESTful web API.

We will present the development of a rich internet application for remote visualization, collaborative annotation, and automated analysis of high-resolution, high-throughput biological images. With our application, high-resolution images in various formats (e.g. from whole-slide scanning devices) that are usually too big to fit into traditional computer memory can be visualized at multiple resolutions in web clients through fully Javascript interfaces, caching mechanisms and distributed image tile servers. Our underlying relational data model allows to create projects which contain users with permission lists, images, ontologies with domainspecific terms, and layers of annotation geometries (e.g. polygons) drawn on top of original images. All project data are stored in a relational database and can be visualized and edited through the web interface, and they can also be retrieved or updated by third-party softwares through a RESTful API (two client libraries in Java and Python have been developed). In addition, we are integrating ImageJ/Fiji/OpenCV image processing routines and our generalpurpose machine learning algorithms to facilitate image annotation and quantification. At this stage of development, our application already delivers about one thousand whole-slide images (roughly 1.5Tb of data) and more than six thousand regions of interest were annotated by our collaborators in the context of ongoing studies on lung cancer and inflammation, and Zebrafish toxicology. Overall, the proposed web software is generally applicable and its methodological choices open the door for large-scale distributed and collaborative image annotation and exploitation projects. Acknowledgments: This work is funded by the research grant n°1017072 of the Walloon Region (DGO6).

automatic exploitation of high-resolution bioimages using tree-based machine learning methods Benjamin Stevens1,2,∗, Loic Rollus1,2,∗, Louis Wehenkel1,2, and Raphael Maree1,2,3 1GIGA-Systems Biology & Chemical Biology, GIGA-R, University of Liege, Belgium 2Systems and Modeling Research Unit, Montefiore Institute, University of Liege, Belgium 3GIGA Bioinformatics Core Facility, GIGA, University of Liege, Belgium These two authors contributed equally to this work.