אירועים והרצאות בפקולטה למדעי המחשב ע"ש הנרי ומרילין טאוב

There is a growing number of tasks that work directly on point clouds. As the size of the point cloud grows, so do the computational demands of these tasks. A possible solution is to sample the point cloud first. Classic sampling approaches, such as farthest point sampling (FPS), do not consider the downstream task. A recent work showed that learning a task-specific sampling can improve results significantly. However, the proposed technique did not deal with the non-differentiability of the sampling operation and offered a workaround instead. We introduce a novel differentiable relaxation for point cloud sampling that approximates sampled points as a mixture of points in the primary input cloud. Our approximation scheme leads to consistently good results on classification and geometry reconstruction applications. We also show that the proposed sampling method can be used as a front to a point cloud registration network. This is a challenging task since sampling must be consistent across two different point clouds for a shared downstream task. In all cases, our approach outperforms existing non-learned and learned sampling alternatives. The talk will cover the following recent papers: 1. Learning to Sample, CVPR 2019. 2. SampleNet: Differentiable Point Cloud Sampling, CVPR 2020 (Oral) Bio: I'm a PhD candidate at Tel Aviv University, advised by Professor Shai Avidan. My current research interests are computer vision and learning methods for 3D point clouds. I received my BSc degree in Electrical Engineering and Physics from the Technion (Summa Cum Laude), as an alumnus of "Psagot", the elite academic program of the IDF. I hold an MSc degree in Electrical Engineering from Tel Aviv University, where I studied lesion segmentation in medical images.