Papers of the day   All papers

PointFlow: 3D Point Cloud Generation with Continuous Normalizing Flows

Comments

Jul 01 2019 Ming-Yu Liu

Check out PointFlow for point cloud generation https://arxiv.org/abs/1906.12320 video http://bit.ly/2ZZydmH code http://bit.ly/2Nr8thP project http://bit.ly/2RMTKML Brought to you by @YangGuandao @xunhuang1995 Zekun Hao @SergeBelongie Bharath Hariharan https://t.co/qiivJkGf67
1 replies, 243 likes


Jul 05 2019 Daisuke Okanohara

PointFlow is a generative model of 3D point clouds. It first generates a latent variable for shape and then uses this to define continuous normalizing flow dynamics and generate point clouds following the dynamics. Training can be done with MLE. https://twitter.com/hillbig/status/1146947411559903232 https://t.co/j4H9rcnRNS
0 replies, 62 likes


Jul 01 2019 Andrew Davison

Another really interesting looking paper on 3D shape representation. Excellent video and animations!
0 replies, 53 likes


Jul 01 2019 Yad Konrad

"Our key insight is that instead of directly parametrizing the distribution of points in a shape, we model this distribution as an invertible parameterized transformation of 3D points from a prior distribution (e.g., a 3D Gaussian)" https://t.co/sWQ0x7sWpC
1 replies, 1 likes


Content