Alessia Atzeni - George Dwyer
May 30, 2018 01:00 PM
End: May 30, 2018 02:00 PM
Location: UCL Bloomsbury - Anatomy Gavin de Beer
Alessia Atzeni - A probabilistic model combining deep learning and multi-atlas segmentation for semi-automated labeling of histology.
Abstract - Thanks to their high resolution and contrast enhanced by different stains, histological images are becoming increasingly widespread in the atlas building literature. Building atlases with histology requires manually delineating a set of regions of interest on a large amount of sections in a stack. This process is not only tedious and time-consuming, but also rather inefficient because the human labeler generally spends much time segmenting adjacent sections that are very similar to one another. Here we propose a probabilistic model for semi-automated segmentation of stacks of histological sections, in which the user manually labels a sparse set of sections (e.g., one every n), and lets the algorithm complete the segmentation in the rest of the stack. The proposed model integrates in a principled manner two families of segmentation techniques that have been very successful in brain imaging: multi-atlas segmentation (MAS) and convolutional neural networks (CNNs). Furthermore, we present a Generalised Expectation Maximisation algorithm to compute the most likely segmentation. Experiments on the Allen dataset show that the model successfully combines the strengths of both techniques (effective label propagation of MAS, and robustness to misregistration of CNNs), and produces significantly more accurate results than using either of them independently.
George Dwyer - Robotically Actuated Endoscopes for Fetal Interventions
Abstract - Fetoscopic interventions require low diameter instrumentation with limited articulation to be introduced into the amniotic sac. Within the amniotic sac, these endoscopes are then used for delicate tasks such as laser photocoagulation of blood vessels on the placenta and the placement of shunts in the fetus' airway. This talk will focus on work to improve the stability and increase the dexterity of endoscopes using different forms of robotics. We present the use of an articulated robot arm to improve the stability of the endoscope and the use of a concentric tube robot to increase the dexterity of the endoscope.