Deep Learning 3D Shape Surfaces Using Geometry Images

By Ayan Sinha, Jing Bai, Karthik Ramani. 2016

336 citations recorded by the source corpus.

Subfields: Computational Mechanics, Computer Graphics and Computer-Aided Design, Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computational Mechanics

In Computer Graphics and Computer-Aided Design

In Computer Vision and Pattern Recognition

Citations in the corpus

Learning Descriptor Networks for 3D Shape Synthesis and Analysis (2018, neutral): “[23] 88.4% PANORAMA-NN [19] 91.1% ECC [22] 90.0% 3D ShapeNets [29] 83.5% DeepPano [21] 85.5% SPH [10] 79.8% VConv-DAE [20] 80.5% 3D-GAN [28] 91.0% 3D DescriptorNet (ours) 92.4 %”
Geometric Deep Learning on Graphs and Manifolds Using Mixture Model CNNs (2016, methodology): “[35] used geometry images representation to obtain a Euclidean parametrization of 3D shapes on which standard CNNs can be applied.”
Generating 3D faces using Convolutional Mesh Autoencoders (2018, background): “Sinha et al. [39] and Maron et al. [30] embed mesh surfaces into planar images to apply conventional CNNs. Sinha et al. use a robust spherical parametrization to project the surface onto an octahedron, which is then cut and unfolded to form a square image.”
DeepSDF: Learning Continuous Signed Distance Functions for Shape Representation (2019, methodology): “Other mesh-based methods use existing [48, 36] or learned [22, 23] parameterization techniques to describe 3D surfaces by morphing 2D planes.”
Learning Implicit Fields for Generative Shape Modeling (2018, background): “There have been a variety of 3D shape representations for deep learning of shapes, such as voxel grids [9, 14, 28, 45, 46], octrees [18, 35, 40, 43, 44], multi-view images [29, 39], point clouds [1, 12, 13, 32, 33, 47, 48], geometry images [37, 38], deformable mesh/patches [16, 38, 42, 47], and part”
A Papier-Mache Approach to Learning 3D Surface Generation (2018, positive): “The geometry image representation [10, 27] views 3D shapes as functions (e.g., vertex positions) embedded in a 2D domain, providing a natural input for 2D neural networks [28].”
SurfaceNet: An End-to-End 3D Neural Network for Multiview Stereopsis (2017, background): “In order to simplify the retrieval and representation of 3D shapes by CNNs, [24] introduces a representation, termed geometry image, which is a 2D representation that approximates a 3D shape.”
Learning a Multi-View Stereo Machine (2017, background): “Approaches which predict shape from a single image have been proposed in form of direct depth map regression [46, 31, 10], generating multiple depth maps from novel viewpoints [51], producing voxel occupancies [5, 16], geometry images [49] and point clouds [11].”
Octree Generating Networks: Efficient Convolutional Architectures for High-resolution 3D Outputs (2017, background): “[30] convert shapes into two-dimensional geometry images and process those with conventional CNNs – an approach only applicable to certain classes of topologies.”
SurfNet: Generating 3D Shape Surfaces Using Deep Residual Networks (2017, methodology): “Next, a randomly chosen mesh model is authalically and spherically parameterized using the method of [28].”
Multi-directional geodesic neural networks via equivariant convolution (2018, methodology): “Another recent method [Ezuz et al. 2017] consists in optimizing an embedding of the shape onto the planar domain using intrinsic metric alignment [Solomon et al. 2016].”
SPLATNet: Sparse Lattice Networks for Point Cloud Processing (2018, methodology): “Another class of methods embeds the input shapes into 2D parametric domains and then execute convolutions within these domains [38, 26, 13].”
HyperCube: Implicit Field Representations of Voxelized 3D Models (2021, background): “3D objects can be represented using different approaches including voxel grids [4, 5, 6, 7], octrees [8, 9, 10, 11], multi-view images [12, 13, 14], point clouds [15, 16, 17, 18, 19], geometry images [20, 21], deformable meshes [5, 21, 22, 19], and part-based structural graphs [23, 24].”
3D Object Classification via Spherical Projections (2017, methodology): “[24] proposes a way to convert 3D objects to geometry images and implicitly learn the geometry and topology of 3D objects.”
GWCNN: A Metric Alignment Layer for Deep Shape Analysis (2017, positive): “We use three methods to map the shape features to an image: the method used by Sinha et al. [SBR16] (Geometry Images) for deep learning (left), GW mapping directly to a 2D grid (middle), and GW mapping based on the learned metric (right).”
General-Purpose Deep Point Cloud Feature Extractor (2018, background): “Another avenue researchers frequently choose is to map the 3D data to 2D, either by creating images from multiple angles [48, 5, 24, 60], or by creating a mapping function from mesh to image such that depth information is preserved [44, 47].”
Learning a Hierarchical Latent-Variable Model of 3D Shapes (2017, methodology): “vised state-of-the-art, outperforming models such as 3D ShapeNet [45], DeepPano [38], and the geometry imagebase approach [39], by a large margin, and comes close to models such as VoxNet [27].”
View Inter-Prediction GAN: Unsupervised Representation Learning for 3D Shapes by Learning Global Shape Memories to Support Local View Predictions (2018, background): “(Bai et al. 2017; Shi et al. 2015; Sﬁkas, Theoharis, and Pratikakis 2017; Sinha, Bai, and Ramani 2016; Su et al. 2015; Johns, Leutenegger, and Davison 2016; Kanezaki, Matsushita, and Nishida 2018), which aims to capture the mapping between 3D raw representations and class labels.”