Representing Point Clouds with Generative Conditional Invertible Flow Networks

In this paper, we propose a simple yet effective method to represent point clouds as sets of samples drawn from a cloud-specific probability distribution. This interpretation matches intrinsic characteristics of point clouds: the number of points and their ordering within a cloud is not important as all points are drawn from the proximity of the object boundary. We postulate to represent each cloud as a parameterized probability distribution defined by a generative neural network. Once trained, such a model provides a natural framework for point cloud manipulation operations, such as aligning a new cloud into a default spatial orientation. To exploit similarities between same-class objects and to improve model performance, we turn to weight sharing: networks that model densities of points belonging to objects in the same family share all parameters with the exception of a small, object-specific embedding vector. We show that these embedding vectors capture semantic relationships between objects. Our method leverages generative invertible flow networks to learn embeddings as well as to generate point clouds. Thanks to this formulation and contrary to similar approaches, we are able to train our model in an end-to-end fashion. As a result, our model offers competitive or superior quantitative results on benchmark datasets, while enabling unprecedented capabilities to perform cloud manipulation tasks, such as point cloud registration and regeneration, by a generative network.

[1]  Chong Xiang,et al.  Generating 3D Adversarial Point Clouds , 2018, 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[2]  Shakir Mohamed,et al.  Variational Inference with Normalizing Flows , 2015, ICML.

[3]  Leonidas J. Guibas,et al.  PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Space , 2017, NIPS.

[4]  Samy Bengio,et al.  Density estimation using Real NVP , 2016, ICLR.

[5]  Alberto L. Sangiovanni-Vincentelli,et al.  A LiDAR Point Cloud Generator: from a Virtual World to Autonomous Driving , 2018, ICMR.

[6]  Yue Wang,et al.  Dynamic Graph CNN for Learning on Point Clouds , 2018, ACM Trans. Graph..

[7]  Piotr Klukowski,et al.  Adversarial autoencoders for compact representations of 3D point clouds , 2018, Comput. Vis. Image Underst..

[8]  Junseok Kwon,et al.  3D Point Cloud Generative Adversarial Network Based on Tree Structured Graph Convolutions , 2019, 2019 IEEE/CVF International Conference on Computer Vision (ICCV).

[9]  Max Welling,et al.  Auto-Encoding Variational Bayes , 2013, ICLR.

[10]  Prafulla Dhariwal,et al.  Glow: Generative Flow with Invertible 1x1 Convolutions , 2018, NeurIPS.

[11]  Christopher M. Bishop,et al.  Pattern Recognition and Machine Learning (Information Science and Statistics) , 2006 .

[12]  Jacek Tabor,et al.  HyperFlow: Representing 3D Objects as Surfaces , 2020, ArXiv.

[13]  Gordon Wetzstein,et al.  Implicit Neural Representations with Periodic Activation Functions , 2020, NeurIPS.

[14]  Yoshua Bengio,et al.  Generative Adversarial Nets , 2014, NIPS.

[15]  Jian Sun,et al.  Deep Residual Learning for Image Recognition , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[16]  Yang Zhang,et al.  Point Cloud GAN , 2018, DGS@ICLR.

[17]  Leonidas J. Guibas,et al.  Frustum PointNets for 3D Object Detection from RGB-D Data , 2017, 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.

[18]  Leonidas J. Guibas,et al.  PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[19]  Leonidas J. Guibas,et al.  Learning Representations and Generative Models for 3D Point Clouds , 2017, ICML.

[20]  Richard S. Zemel,et al.  Generative Moment Matching Networks , 2015, ICML.

[21]  Alexandre Boulch,et al.  Unstructured Point Cloud Semantic Labeling Using Deep Segmentation Networks , 2017, 3DOR@Eurographics.

[22]  T. Trzciński,et al.  Hypernetwork approach to generating point clouds , 2020, ICML.

[23]  Yoshua Bengio,et al.  Extracting and composing robust features with denoising autoencoders , 2008, ICML '08.

[24]  Nikolaus Hansen,et al.  The CMA Evolution Strategy: A Comparing Review , 2006, Towards a New Evolutionary Computation.

[25]  Ming-Wei Chang,et al.  BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding , 2019, NAACL.

[26]  David Duvenaud,et al.  Neural Ordinary Differential Equations , 2018, NeurIPS.

[27]  Bin Yang,et al.  PIXOR: Real-time 3D Object Detection from Point Clouds , 2018, 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.

[28]  Jianxiong Xiao,et al.  3D ShapeNets: A deep representation for volumetric shapes , 2014, 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[29]  Ming-Yu Liu,et al.  PointFlow: 3D Point Cloud Generation With Continuous Normalizing Flows , 2019, 2019 IEEE/CVF International Conference on Computer Vision (ICCV).

[30]  Jiajun Wu,et al.  Learning a Probabilistic Latent Space of Object Shapes via 3D Generative-Adversarial Modeling , 2016, NIPS.

[31]  Kurt Keutzer,et al.  SqueezeSeg: Convolutional Neural Nets with Recurrent CRF for Real-Time Road-Object Segmentation from 3D LiDAR Point Cloud , 2017, 2018 IEEE International Conference on Robotics and Automation (ICRA).

[32]  Pieter Abbeel,et al.  Image Object Label 3 D CAD Model Candidate Grasps Google Object Recognition Engine Google Cloud Storage Select Feasible Grasp with Highest Success Probability Pose EstimationCamera Robots Cloud 3 D Sensor , 2014 .