- 31
- 62 140
UofT CSC 2547 3D & Geometric Deep Learning
Canada
Приєднався 10 лют 2021
This course introduces deep learning methods and modern advances in 3D Vision. We will study representations, learning algorithms, and generative models for 3D vision tasks at the object and scene level. We will then study Geometric Deep Learning and concepts of Manifold Learning as relevant to Deep Learning. The 3D nature of this topic has many potential applications in graphics, robotics, content creation, mixed reality, biometrics, and more.
This channel will have ~10 min tutorials on the top papers in the field.
This channel will have ~10 min tutorials on the top papers in the field.
CSC2547 Learning Delaunay Surface Elements for Mesh Reconstruction
Title:
Learning Delaunay Surface Elements for Mesh Reconstruction
Authors:
Marie-Julie Rakotosaona, Paul Guerrero, Noam Aigerman, Niloy Mitra, Maks Ovsjanikov
Presenter:
Brendan Duke
Learning Delaunay Surface Elements for Mesh Reconstruction
Authors:
Marie-Julie Rakotosaona, Paul Guerrero, Noam Aigerman, Niloy Mitra, Maks Ovsjanikov
Presenter:
Brendan Duke
Переглядів: 1 151
Відео
CSC2547 Isometric Transformation Invariant and Equivariant Graph Convolutional Networks
Переглядів 1,1 тис.3 роки тому
Title: Isometric Transformation Invariant and Equivariant Graph Convolutional Networks Authors: Masanobu Horie, Naoki Morita, Toshiaki Hishinuma, Yu Ihara, Naoto Mitsume Presentor: Sejin Kim
CSC2547 LanczosNet: Multi-Scale Deep Graph Convolutional Networks
Переглядів 4283 роки тому
Title: LanczosNet: Multi-Scale Deep Graph Convolutional Networks Authors: Renjie Liao, Zhizhen Zhao, Raquel Urtasun, Richard S. Zemel Presentor: Juan Carrillo
CSC2547 Relational inductive biases, deep learning, and graph networks
Переглядів 2,4 тис.3 роки тому
Title: Relational inductive biases, deep learning, and graph networks Authors: Peter W. Battaglia, Jessica B. Hamrick, Victor Bapst,Alvaro Sanchez-Gonzalez, Vinicius Zambaldi, Mateusz MalinowskiAndrea Tacchetti, David Raposo, Adam Santoro, Ryan Faulkner,Caglar Gulcehre, Francis Song, Andrew Ballard, Justin Gilmer,George Dahl, Ashish Vaswani, Kelsey Allen, Charles Nash,Victoria Langston, Chris D...
CSC2547 SIREN: Implicit Neural Representations with Periodic Activation Functions
Переглядів 2,9 тис.3 роки тому
Title: Implicit Neural Representations with Periodic Activation Functions Authors: Vincent Sitzmann*, Julien N. P. Martel*, Alexander Bergman, David B. Lindell, Gordon Wetzstein Presentor: Zikun Chen
CSC2547 Fast end-to-end learning on protein surfaces
Переглядів 8223 роки тому
Title: Fast end-to-end learning on protein surfaces Authors: Freyr Sverrisson, Jean Feydy, Bruno E. Correia, Michael M. Bronstein
CSC2547 NeMo: Neural Mesh Models of Contrastive Features for Robust 3D Pose Estimation
Переглядів 6543 роки тому
Title: NeMo: Neural Mesh Models of Contrastive Features for Robust 3D Pose Estimation Authors: Angtian Wang, Adam Kortylewski, Alan Yuille Presentor: William Ngo
csc2547 Canonical Capsules: Unsupervised Capsules in Canonical Pose
Переглядів 1763 роки тому
Title: Canonical Capsules: Unsupervised Capsules in Canonical Pose Authors: Weiwei Sun, Andrea Tagliasacchi, Boyang Deng, Sara Sabour, Soroosh Yazdani, Geoffrey Hinton, Kwang Moo Yi Presentor: Ioannis Xarchakos
CSC2547 On learning sets of symmetric elements
Переглядів 2363 роки тому
Title On Learning Sets of Symmetric Elements Authors: Haggai Maron, Or Litany, Gal Chechik, Ethan Fetaya Presentor: Dmitrii Shubin
CSC2547 Unsupervised Learning of Probably Symmetric Deformable 3D Objects from Images in the Wild
Переглядів 3603 роки тому
Title: Unsupervised Learning of Probably Symmetric Deformable 3D Objects from Images in the Wild Authors: Shangzhe Wu, Christian Rupprecht, Andrea Vedaldi Presentor: Brendan Kolisnik
CSC2547 CNNs on Surfaces using Rotation Equivariant Features
Переглядів 4183 роки тому
Title: CNNs on Surfaces using Rotation-Equivariant Features Authors: Ruben Wiersma, Elmar Eisemann, and Klaus Hildebrandt Presentor: Shichen Lu
CSC2547 iNeRF Inverting Neural Radiance Fields for Pose Estimation
Переглядів 6933 роки тому
Title: iNeRF: Inverting Neural Radiance Fields for Pose Estimation Authors: Lin Yen-Chen, Pete Florence, Jonathan T. Barron, Alberto Rodriguez, Phillip Isola, Tsung-Yi Lin Presentor: Bin Shi
CSC2547 Gauge Equivariant Mesh CNNs: Anisotropic convolutions on geometric graphs
Переглядів 1,4 тис.3 роки тому
Title: Gauge Equivariant Mesh CNNs: Anisotropic convolutions on geometric graphs Author: Pim De Haan, Maurice Weiler, Taco Cohen, Max Welling Presentor: Otman Benchekroun
CSC2547 Deformable Neural Radiance Fields
Переглядів 7173 роки тому
Paper Title: Deformable Neural Radiance Fields Author: Keunhong Park, Utkarsh Sinha, Jonathan T. Barron, Sofien Bouaziz, Dan B Goldman, Steven M. Seitz, Ricardo Martin-Brualla Presentor: Yun-Chun Chen
CSC2547 Neural Sparse Voxel Fields
Переглядів 1,5 тис.3 роки тому
Paper Title: Neural Sparse Voxel Fields Authors: Lingjie Liu, Jiatao Gu, Kyaw Zaw Lin, Tat-Seng Chua, Christian Theobalt Presentor: Tianxing Li
CSC2547 Neural Reflectance Fields for Appearance Acquisition
Переглядів 1 тис.3 роки тому
CSC2547 Neural Reflectance Fields for Appearance Acquisition
CSC2547 PolyGen: An Autoregressive Generative Model of 3D Meshes
Переглядів 2,8 тис.3 роки тому
CSC2547 PolyGen: An Autoregressive Generative Model of 3D Meshes
CSC2547 Virtual Multi-view Fusion for 3D Semantic Segmentation
Переглядів 2733 роки тому
CSC2547 Virtual Multi-view Fusion for 3D Semantic Segmentation
CSC2547 NeRF in the Wild Neural Radiance Fields for Unconstrained Photo Collections
Переглядів 2 тис.3 роки тому
CSC2547 NeRF in the Wild Neural Radiance Fields for Unconstrained Photo Collections
CSC2547 Learning Generative Models of 3D Structures
Переглядів 8053 роки тому
CSC2547 Learning Generative Models of 3D Structures
Differentiable Volumetric Rendering: Learning Implicit 3D Representations without 3D Supervision
Переглядів 1,8 тис.3 роки тому
Differentiable Volumetric Rendering: Learning Implicit 3D Representations without 3D Supervision
CSC2547 Scene Representation Network Continuous 3D-Structure-Aware Neural Scene Representations
Переглядів 7093 роки тому
CSC2547 Scene Representation Network Continuous 3D-Structure-Aware Neural Scene Representations
CSC2547 Multiview Neural Surface Reconstruction by Disentangling Geometry and Appearance
Переглядів 1,9 тис.3 роки тому
CSC2547 Multiview Neural Surface Reconstruction by Disentangling Geometry and Appearance
CSC2547 Differentiable Monte Carlo Ray Tracing through Edge Sampling
Переглядів 1,9 тис.3 роки тому
CSC2547 Differentiable Monte Carlo Ray Tracing through Edge Sampling
CSC2547 Differentiable Rendering A Survey
Переглядів 6 тис.3 роки тому
CSC2547 Differentiable Rendering A Survey
CSC2547 Dynamic Graph CNN for Learning on Point Cloud
Переглядів 3,5 тис.3 роки тому
CSC2547 Dynamic Graph CNN for Learning on Point Cloud
CSC2547 Learning Gradient Fields for Shape Generation
Переглядів 3223 роки тому
CSC2547 Learning Gradient Fields for Shape Generation
CSC2547 ShapeAssembly Learning to Generate Programs for 3D Shape Structure Synthesis
Переглядів 4873 роки тому
CSC2547 ShapeAssembly Learning to Generate Programs for 3D Shape Structure Synthesis
CSC2547 KPConv Flexible and Deformable Convolution for Point Clouds
Переглядів 4,6 тис.3 роки тому
CSC2547 KPConv Flexible and Deformable Convolution for Point Clouds
CSC2547 DeepSDF Learning Continuous Signed Distance Functions for Shape Representation
Переглядів 19 тис.3 роки тому
CSC2547 DeepSDF Learning Continuous Signed Distance Functions for Shape Representation
very good video. let me know about graph network.
Quitzon Fields
What is the latest work on DeepSDF, any update?
Is the code available?
Great Explanation!
Your explanations are amazing
沙发
Are multiple rays emitted from the same pixel parallel to each other?
Awesome work :)
thank you for the tutorial. Can point cloud represents small objects too, like in the case of CAD data, there are small holes that need to be detected. So can point cloud does that. And how many points can the data be converted into points? Is the max point is 4096 ?
great work
Thank you for your effort and explanation, but i have an objection, authors of the paper state that using rigid neighbourhoods are better for point clouds with variable densities i.e. lidar scans, since they tend to keep spatial correlation zones more locally and not extending towards other further points. I also have seen a similar statement in Ψ-CNN paper.
great talk! thanks
Can I get a copy of the slides ？plz
Thank you very much! Your explanation has deepened my understanding of this paper.
thank you so much , this was a great explanation
Clear explanation
This channel is seriously underrated. Wonderful videos. Btw, I am 3D artist learning mathematics and physics. I use Blender.
me too
Great tutorial. Thank you. I just subscribed.