Filters








2,763 Hits in 3.2 sec

Geometric Disentanglement for Generative Latent Shape Models [article]

Tristan Aumentado-Armstrong, Stavros Tsogkas, Allan Jepson, Sven Dickinson
2019 arXiv   pre-print
One avenue that has recently begun to be explored is the use of latent representations of generative models.  ...  In this paper, we propose an unsupervised approach to partitioning the latent space of a variational autoencoder for 3D point clouds in a natural way, using only geometric information.  ...  Acknowledgments We are grateful for support from NSERC (CGS-M-510941-2017) and Samsung Research.  ... 
arXiv:1908.06386v1 fatcat:dtjbjhjxz5dlnjkth5f3o4tkaa

Deformable Generator Network: Unsupervised Disentanglement of Appearance and Geometry [article]

Xianglei Xing, Ruiqi Gao, Tian Han, Song-Chun Zhu, Ying Nian Wu
2020 arXiv   pre-print
We present a deformable generator model to disentangle the appearance and geometric information for both image and video data in a purely unsupervised manner.  ...  Two generators take independent latent vectors as input to disentangle the appearance and geometric information from image or video sequences.  ...  We also introduce a dynamic deformable generator model for the spatial-temporal process which disentangle the appearance and geometric information of a video sequence into two groups of independent latent  ... 
arXiv:1806.06298v3 fatcat:cwx4l5crqjhsfagzctvoylvivq

Unsupervised Disentangling of Appearance and Geometry by Deformable Generator Network

Xianglei Xing, Tian Han, Ruiqi Gao, Song-Chun Zhu, Ying Nian Wu
2019 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)  
Two generators act upon independent latent factors to extract disentangled appearance and geometric information from images.  ...  We present a deformable generator model to disentangle the appearance and geometric information in purely unsupervised manner.  ...  The model can be generalized to model dynamic patterns by adding transition models for the latent vectors.  ... 
doi:10.1109/cvpr.2019.01060 dblp:conf/cvpr/XingHGZW19 fatcat:q4shcfvdrnedtnxgn6ocsl2gde

DSG-Net: Learning Disentangled Structure and Geometry for 3D Shape Generation [article]

Jie Yang, Kaichun Mo, Yu-Kun Lai, Leonidas J. Guibas, Lin Gao
2021 arXiv   pre-print
To tackle this, we introduce DSG-Net, a deep neural network that learns a disentangled structured and geometric mesh representation for 3D shapes, where two key aspects of shapes, geometry, and structure  ...  While significant progress has been made, especially with recent deep generative models, it remains a challenge to synthesize high-quality shapes with rich geometric details and complex structure, in a  ...  G MORE RESULTS ON DISENTANGLED SHAPE GENERATION DSG-Net learns two disentangled latent spaces for modeling shape structure and geometry respectively, which enables a novel application of disentangled shape  ... 
arXiv:2008.05440v3 fatcat:gth6s7jmh5fbjogehj72o3exoe

Intrinsic-Extrinsic Preserved GANs for Unsupervised 3D Pose Transfer [article]

Haoyu Chen, Hao Tang, Henglin Shi, Wei Peng, Nicu Sebe, Guoying Zhao
2021 arXiv   pre-print
., the need for manually encoding for shape and pose disentanglement. In this paper, we present an unsupervised approach to conduct the pose transfer between any arbitrate given 3D meshes.  ...  Specifically, a novel Intrinsic-Extrinsic Preserved Generative Adversarial Network (IEP-GAN) is presented for both intrinsic (i.e., shape) and extrinsic (i.e., pose) information preservation.  ...  As well, the authors wish to acknowledge CSC-IT Center for Science, Finland, for computational resources.  ... 
arXiv:2108.07520v2 fatcat:ephakoeqd5aprpwdrhb6mbcyia

LIMP: Learning Latent Shape Representations with Metric Preservation Priors [article]

Luca Cosmo, Antonio Norelli, Oshri Halimi, Ron Kimmel, Emanuele Rodolà
2020 arXiv   pre-print
The effectiveness and potential of our generative model is showcased in applications of style transfer, content generation, and shape completion.  ...  In this paper, we advocate the adoption of metric preservation as a powerful prior for learning latent representations of deformable 3D shapes.  ...  More recently, a geometric disentanglement model for deformable point clouds was introduced in [2] .  ... 
arXiv:2003.12283v2 fatcat:bcv2elssh5brhcxtutv52v5fki

DSG-Net: Learning Disentangled Structure and Geometry for 3D Shape Generation

Jie Yang, Kaichun Mo, Yu-Kun Lai, Leonidas J. Guibas, Lin Gao
2022 ACM Transactions on Graphics  
To tackle this, we introduce DSG-Net, a deep neural network that learns a disentangled structured & geometric mesh representation for 3D shapes, where two key aspects of shapes, geometry and structure,  ...  While significant progress has been made, especially with recent deep generative models, it remains a challenge to synthesize high-quality shapes with rich geometric details and complex structures, in  ...  DSG-Net learns two disentangled latent spaces for modeling shape structure and geometry, which enables a novel task of generating shapes with a given shape structure or geometry pattern.  ... 
doi:10.1145/3526212 fatcat:l54jkinxkbcl7jfbsmeabrnf5e

Unsupervised Geometric Disentanglement for Surfaces via CFAN-VAE [article]

N. Joseph Tatro, Stefan C. Schonsheck, Rongjie Lai
2020 arXiv   pre-print
Geometric disentanglement, the separation of latent codes for intrinsic (i.e. identity) and extrinsic(i.e. pose) geometry, is a prominent task for generative models of non-Euclidean data such as 3D deformable  ...  It provides greater interpretability of the latent space, and leads to more control in generation.  ...  Part of this research was performed while the authors were visiting the Institute for Pure and Applied Mathematics (IPAM), which is supported by the NSF DMS-1440415, during the Geometry and Learning from  ... 
arXiv:2005.11622v2 fatcat:rukr2xt3gbam5bvlqfi6oqgtie

Latent feature disentanglement for 3D meshes [article]

Jake Levinson, Avneesh Sud, Ameesh Makadia
2019 arXiv   pre-print
We introduce a supervised generative 3D mesh model that disentangles the latent shape representation into independent generative factors.  ...  Generative modeling of 3D shapes has become an important problem due to its relevance to many applications across Computer Vision, Graphics, and VR.  ...  In this work, we introduce a new generative model for 3D shapes that explicitly disentangles the shape representation by its observable generative factors.  ... 
arXiv:1906.03281v1 fatcat:paghb5ku2bds3mwm5q5hfdvzq4

3D Shape Variational Autoencoder Latent Disentanglement via Mini-Batch Feature Swapping for Bodies and Faces [article]

Simone Foti, Bongjin Koo, Danail Stoyanov, Matthew J. Clarkson
2022 arXiv   pre-print
Learning a disentangled, interpretable, and structured latent representation in 3D generative models of faces and bodies is still an open problem.  ...  Experimental results conducted on 3D meshes show that state-of-the-art methods for latent disentanglement are not able to disentangle identity features of faces and bodies.  ...  Autoencoder Latent Disentanglement Latent disentanglement for the generation of 3D shapes has been explored mostly in relation to the disentanglement of identity and pose generative factors. [3] created  ... 
arXiv:2111.12448v5 fatcat:xtpexy7h5faazbd4tm63q7hbom

Disentangling Geometric Deformation Spaces in Generative Latent Shape Models [article]

Tristan Aumentado-Armstrong, Stavros Tsogkas, Sven Dickinson, Allan Jepson
2021 arXiv   pre-print
In this work, we improve on a prior generative model of geometric disentanglement for 3D shapes, wherein the space of object geometry is factorized into rigid orientation, non-rigid pose, and intrinsic  ...  We evaluate our model on its generative modelling, representation learning, and disentanglement performance, showing improved rotation invariance and intrinsic-extrinsic factorization quality over the  ...  Latent Variational Autoencoder Model Overview Our goal is to define a disentangled generative model of 3D shapes, using a VAE.  ... 
arXiv:2103.00142v1 fatcat:oefxuozl5bewtcnbvqoe66ccze

Geometry of Deep Generative Models for Disentangled Representations [article]

Ankita Shukla, Shagun Uppal, Sarthak Bhagat, Saket Anand, Pavan Turaga
2019 arXiv   pre-print
In this work, we explore the geometry of popular generative models for disentangled representation learning.  ...  In new developments, deep generative models have been used for learning semantically meaningful 'disentangled' representations; that capture task relevant attributes while being invariant to other attributes  ...  ACKNOWLEDGEMENT This work is supported in part by Infosys Center for Artificial intelligence at IIIT Delhi, India.  ... 
arXiv:1902.06964v1 fatcat:ttyd7v3hqvgq7o35ydillvx5aa

Disentangling Geometry and Appearance with Regularised Geometry-Aware Generative Adversarial Networks

Linh Tran, Jean Kossaifi, Yannis Panagakis, Maja Pantic
2019 International Journal of Computer Vision  
In this work, we propose a regularized Geometry-Aware Generative Adversarial Network (GAGAN) which disentangles appearance and shape in the latent space.  ...  However, currently available generative models do not incorporate geometric information into the image generation process. This often yields visual objects of degenerated quality.  ...  In particular, we introduce Geometry-Aware GAN (GAGAN) which disentangles the latent space corresponding to shape and texture by employing a statistical shape model.  ... 
doi:10.1007/s11263-019-01155-7 fatcat:4tdlc36yzvgu3cmc7iq7fbcdiq

MRGAN: Multi-Rooted 3D Shape Generation with Unsupervised Part Disentanglement [article]

Rinon Gal, Amit Bermano, Hao Zhang, Daniel Cohen-Or
2020 arXiv   pre-print
We demonstrate the controllability offered by our part-disentangled generation through two applications for shape modeling: part mixing and individual part variation, without receiving segmented shapes  ...  We present MRGAN, a multi-rooted adversarial network which generates part-disentangled 3D point-cloud shapes without part-based shape supervision.  ...  Our method falls firmly into the camp of networks designed for disentangling over pre-determined, yet latent, properties, i.e., the geometric and structural properties of shape parts.  ... 
arXiv:2007.12944v1 fatcat:7twrpqwk5feftkughbzykwui4m

Guided Variational Autoencoder for Disentanglement Learning [article]

Zheng Ding, Yifan Xu, Weijian Xu, Gaurav Parmar, Yang Yang, Max Welling, Zhuowen Tu
2020 arXiv   pre-print
We propose an algorithm, guided variational autoencoder (Guided-VAE), that is able to learn a controllable generative model by performing latent representation disentanglement learning.  ...  Guided-VAE enjoys its transparency and simplicity for the general representation learning task, as well as disentanglement learning.  ...  ZD is supported by the Tsinghua Academic Fund for Undergraduate Overseas Studies. We thank Kwonjoon Lee, Justin Lazarow, and Jilei Hou for valuable feedbacks.  ... 
arXiv:2004.01255v1 fatcat:ghdtml3harb4rdriohnwnyv32y
« Previous Showing results 1 — 15 out of 2,763 results