Computer Vision News - June 2021

24 Challenge of the Month and it will be a major contribution to the community. It has been reviewed by expert surgeons and features 18 fetoscopy videos annotated for placental vessel segmentation . They will also be releasing sequences from these videos for registration and mosaicking, which means creating an image of the increased field of view that can be considered a panorama image inside the human cavity. It will be the first large-scale multi- center fetoscopy dataset. Three fetal medical centers have contributed: University College London Hospital; Istituto Giannina Gaslini in Genoa, Italy; and University Hospital Leuven in Belgium. Sara is an Assistant Professor at the BioRobotics Institute of Scuola Superiore Sant’Anna in Italy . Her main research interests are in surgical data science and deep learning methodologies applied to the analysis of medical images and videos. “If you develop a deep learning algorithm that works well for one surgeon, or one endoscope, or one center, that isn’t particularly useful as it creates a bias problem,” she explains. “We are so proud to share this multi- center dataset , which is something new and will pave the way for future challenges.” They aim to bring this technology into a real operating room within the next couple of years. Finally, we asked the team to tell us what fascinates them most about working in this field. “The most magical part of our work is you never know what ideas other researchers may have based on what you have done,” Sara smiles. “You do