1,885 Hits in 3.6 sec

A deep perceptual metric for 3D point clouds [article]

Maurice Quach, Aladine Chetouani, Giuseppe Valenzise, Frederic Dufaux
2021 arXiv   pre-print
We thus propose a perceptual loss function for 3D point clouds which outperforms existing loss functions on the ICIP2020 subjective dataset.  ...  Point clouds are essential for storage and transmission of 3D content. As they can entail significant volumes of data, point cloud compression is crucial for practical usage.  ...  Acknowledgments We would like to thank the authors of [17] for providing their implementation of naBCE. This work was funded by the ANR ReVeRy national fund (REVERY ANR-17-CE23-0020).  ... 
arXiv:2102.12839v1 fatcat:x3rj2wouifajphgfnzzpc7l5iy

Pixel2Mesh: Generating 3D Mesh Models from Single RGB Images [article]

Nanyang Wang, Yinda Zhang, Zhuwen Li, Yanwei Fu, Wei Liu, Yu-Gang Jiang
2018 arXiv   pre-print
Limited by the nature of deep neural network, previous methods usually represent a 3D shape in volume or point cloud, and it is non-trivial to convert them to the more ready-to-use mesh model.  ...  We propose an end-to-end deep learning architecture that produces a 3D shape in triangular mesh from a single color image.  ...  Since the metrics are defined on point cloud, we can evaluate PSG directly on its output, our method by uniformly sampling point on surface, and 3D-R2N2 by uniformly sampling point from mesh created using  ... 
arXiv:1804.01654v1 fatcat:oeinb2qx7vgvbmeynz6olt2yp4

Grasping Field: Learning Implicit Representations for Human Grasps [article]

Korrawe Karunratanakul, Jinlong Yang, Yan Zhang, Michael Black, Krikamol Muandet, Siyu Tang
2020 arXiv   pre-print
Specifically, our generative model is able to synthesize high-quality human grasps, given only on a 3D object point cloud.  ...  We name this 3D to 2D mapping as Grasping Field, parameterize it with a deep neural network, and learn it from data.  ...  While MJB is a part-time employee of Amazon, his research was performed solely at MPI. He is an investor in Meshcapde GmbH.  ... 
arXiv:2008.04451v3 fatcat:ytq2g3mb4fhyvlpm7u34z3mdcy

No-Reference Point Cloud Quality Assessment via Domain Adaptation [article]

Qi Yang, Yipeng Liu, Siheng Chen, Yiling Xu, Jun Sun
2021 arXiv   pre-print
We present a novel no-reference quality assessment metric, the image transferred point cloud quality assessment (IT-PCQA), for 3D point clouds.  ...  For quality assessment, deep neural network (DNN) has shown compelling performance on no-reference metric design.  ...  Predicting the perceptual quality of point cloud: A 3d-to-2d projection-based exploration. IEEE Trans.  ... 
arXiv:2112.02851v1 fatcat:qsnxxgy4yve6zk2rnxb2eoglmm

Grasping Field: Learning Implicit Representations for Human Grasps

Korrawe Karunratanakul, Jinlong Yang, Yan Zhang, Michael J. Black, Krikamol Muandet, Siyu Tang
2020 2020 International Conference on 3D Vision (3DV)  
Specifically, our generative model is able to synthesize high-quality human grasps, given only on a 3D object point cloud.  ...  Furthermore, based on the grasping field representation, we propose a deep network for the challenging task of 3D hand-object interaction reconstruction from a single RGB image.  ...  While MJB is a part-time employee of Amazon, his research was performed solely at MPI. He is an investor in Meshcapde GmbH.  ... 
doi:10.1109/3dv50981.2020.00043 fatcat:tki4tvss6nd55ka7malh33bzne

Revisiting Point Cloud Simplification: A Learnable Feature Preserving Approach [article]

Rolandos Alexandros Potamias and Giorgos Bouritsas and Stefanos Zafeiriou
2021 arXiv   pre-print
The approach is extensively evaluated on various datasets using several perceptual metrics.  ...  The recent advances in 3D sensing technology have made possible the capture of point clouds in significantly high resolution.  ...  Several 3D perceptual metric studies (Lee et al, 2005; Lavoué et al, 2006; Lavoué, 2009; Zhang et al, 2019) have pointed out that features such as curvature and roughness of a 3D model are highly correlated  ... 
arXiv:2109.14982v1 fatcat:cgvxhld4z5fafavfau6gt73qni

Editorial: Introduction to the Issue on Deep Learning for Image/Video Restoration and Compression

A. Murat Tekalp, Michele Covell, Radu Timofte, Chao Dong
2021 IEEE Journal on Selected Topics in Signal Processing  
It presents a novel deep learning-based solution for point cloud geometry coding that divides the point cloud into 3D blocks and selects the most suitable available deep learning coding model to code each  ...  The paper entitled "Learning robust graph-convolutional representations for point cloud denoising" proposes a deep learning method that can simultaneously denoise a point cloud and remove outliers in a  ... 
doi:10.1109/jstsp.2021.3053364 fatcat:hjo5pvw6lvgpfga2wfq4vpaq3q

ShapeAdv: Generating Shape-Aware Adversarial 3D Point Clouds [article]

Kibok Lee, Zhuoyuan Chen, Xinchen Yan, Raquel Urtasun, Ersin Yumer
2020 arXiv   pre-print
Our shape-aware adversarial attacks are orthogonal to existing point cloud based attacks and shed light on the vulnerability of 3D deep neural networks.  ...  We develop shape-aware adversarial 3D point cloud attacks by leveraging the learned latent space of a point cloud auto-encoder where the adversarial noise is applied in the latent space.  ...  [1] proposed a two-step training method for an auto-encoder, for learning latent representation space of 3D point clouds and generating 3D point clouds from the latent representation.  ... 
arXiv:2005.11626v1 fatcat:derfn55asneibdpoyblzaoeldy

Neural Point Cloud Rendering via Multi-Plane Projection [article]

Peng Dai, Yinda Zhang, Zhuwen Li, Shuaicheng Liu, Bing Zeng
2020 arXiv   pre-print
We present a new deep point cloud rendering pipeline through multi-plane projections.  ...  Moreover, our pipeline is robust to noisy and relatively sparse point cloud for a variety of challenging scenes.  ...  To recap, we propose a deep learning based method to render images from point cloud.  ... 
arXiv:1912.04645v2 fatcat:qg6ogniob5brvmxwuni7jgdxca

3DSNet: Unsupervised Shape-to-Shape 3D Style Transfer [article]

Mattia Segu, Margarita Grinvald, Roland Siegwart, Federico Tombari
2021 arXiv   pre-print
The proposed method can synthesize new 3D shapes both in the form of point clouds and meshes, combining the content and style of a source and target 3D model to generate a novel shape that resembles in  ...  To our knowledge, we propose the first learning-based approach for style transfer between 3D objects based on disentangled content and style representations.  ...  To compute a perceptual distance d i,j between two point clouds x i and x j given a pre-trained network F , we introduce the 3D-LPIPS metric.  ... 
arXiv:2011.13388v4 fatcat:lgiupah2kvg6hfgeik5fgb5gge

VCIP 2020 Index

2020 2020 IEEE International Conference on Visual Communications and Image Processing (VCIP)  
Metric for Image Quality Assessment He, Yingshen A point cloud compression framework via spherical projection Herglotz, Christian FishUI: Interactive Fisheye Distortion Visualization Herglotz  ...  Shao, Jinxin Orthogonal Features Fusion Network for Anomaly Detection Shao, Yiting Fast Recolor Prediction Scheme in Point Cloud Attribute Compression Shao, Yiting A point cloud compression  ... 
doi:10.1109/vcip49819.2020.9301896 fatcat:bdh7cuvstzgrbaztnahjdp5s5y

Associate-3Ddet: Perceptual-to-Conceptual Association for 3D Point Cloud Object Detection [article]

Liang Du and Xiaoqing Ye and Xiao Tan and Jianfeng Feng and Zhenbo Xu and Errui Ding and Shilei Wen
2020 arXiv   pre-print
Object detection from 3D point clouds remains a challenging task, though recent studies pushed the envelope with the deep learning techniques.  ...  Owing to the severe spatial occlusion and inherent variance of point density with the distance to sensors, appearance of a same object varies a lot in point cloud data.  ...  In [45] , the author first introduced the VoxelNet architecture to learn discriminative features from 3D point clouds for 3D object detection.  ... 
arXiv:2006.04356v1 fatcat:f6e4wjx365hyndi4aztnmqqzfi

Identifying Unknown Instances for Autonomous Driving [article]

Kelvin Wong, Shenlong Wang, Mengye Ren, Ming Liang, Raquel Urtasun
2019 arXiv   pre-print
In this paper, we develop a novel open-set instance segmentation algorithm for point clouds which can segment objects from both known and unknown classes in a holistic way.  ...  Our method uses a deep convolutional neural network to project points into a category-agnostic embedding space in which they can be clustered into instances irrespective of their semantics.  ...  For 3D point cloud, [31, 32] also use similar two-stage architectures to perform point cloud detection and segmentation.  ... 
arXiv:1910.11296v1 fatcat:hcur73y3xrdyfjxsw6zv3hdbye

Rethinking Sampling in 3D Point Cloud Generative Adversarial Networks [article]

He Wang, Zetian Jiang, Li Yi, Kaichun Mo, Hao Su, Leonidas J. Guibas
2020 arXiv   pre-print
We further study how different evaluation metrics weigh the sampling pattern against the geometry and propose several perceptual metrics forming a sampling spectrum of metrics.  ...  Guided by the proposed sampling spectrum, we discover a middle-point sampling-aware baseline discriminator, PointNet-Mix, which improves all existing point cloud generators by a large margin on sampling-related  ...  Generating 3D point clouds with GANs in an unsupervised manner is an important but less explored problem. 3D point cloud GAN learns to transform a random latent code into a 3D surface point cloud by playing  ... 
arXiv:2006.07029v1 fatcat:x3qjxbnqmngfnfoqjkbx4wssdy

Inferring Point Cloud Quality via Graph Similarity [article]

Qi Yang, Zhan Ma, Yiling Xu, Zhu Li, Jun Sun
2020 arXiv   pre-print
We propose the GraphSIM -- an objective metric to accurately predict the subjective quality of point cloud with superimposed geometry and color impairments.  ...  Our GraphSIM is validated using two large and independent point cloud assessment datasets that involve a wide range of impairments (e.g., re-sampling, compression, additive noise), reliably demonstrating  ...  INTRODUCTION With the advancements of 3D capturing and rendering technologies [1] , point cloud has emerged as a promising format for representing 3D object and scene realistically [2] .  ... 
arXiv:2006.00497v2 fatcat:j6724iqrzjdyfgn3pqh4evwzqq
« Previous Showing results 1 — 15 out of 1,885 results