Sophia Bano - Beatrice Van Amsterdam - CMIC/WEISS joint seminar series

Sophia Bano

Title: Deep Learning for Understanding Fetoscopic Videos

Abstract: Fetoscopic laser photocoagulation is a minimally invasive surgery for the treatment of Twin-to-twin Transfusion Syndrome (TTTS). By using a lens/fibre-optic scope, inserted into the amniotic cavity, the abnormal placental vascular anastomoses are identified and ablated to regulate blood flow to both fetuses. Limited field-of-view (FoV), occlusions due to fetus presence and low visibility make it difficult to identify all vascular anastomoses. Computer-assisted techniques can provide better understanding of the anatomical structure and can help the surgeon in navigating through the placenta by creating a larger field-of-view display, identifying vessels, and detecting occlusion-free views. This talk will overview the challenges involved in analyzing fetoscopic videos and our deep learning-based contributions for advancing towards computer-assisted fetoscopy.

Website: https://sophiabano.github.io

Beatrice Van Amsterdam

Title: Surgical gesture segmentation in robot-assisted laparoscopic surgery

Abstract: Robot-Assisted Minimally Invasive Surgery (RAMIS) is now an established practice across a range of surgical specialties, which helps to improve the precision of surgical manipulation and the ergonomic comfort of the surgeon during laparoscopic surgery. With RAMIS a large dataset of video and kinematic signals can be recorded from the robotic platform during an intervention, which can be used for a range of purposes such as surgical skill assessment and automation. However, all of these functionalities are difficult to implement robustly due to the complexity and variability of surgical tasks. A number of researchers addressed the situation by decomposing surgical demonstrations into action-units that characterise specific surgical tasks. Automatic segmentation of these demonstrations into meaningful units can indeed help to develop new metrics for surgical skill assessment as well as to simplify surgical automation.