DualConvMesh-Net: Joint Geodesic and Euclidean Convolutions on 3D Meshes

Schult, Jonas; Engelmann, Francis; Kontogianni, Theodora; Leibe, Bastian

doi:10.1109/cvpr42600.2020.00864

Cited by 86 publications

(61 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In Masci et al [2015] and Monti et al [2017], local coordinate systems are defined to confine the convolution operations on regular grids. MeshCNN [Hanocka et al 2019] executes convolution or pooling on mesh edges while Schult et al [2020] performs graph convolution and vertex pooling on mesh vertices. In Feng et al [2019], introduce a convolution on mesh facets and separate mesh features into spatial and structure levels manually.…”

Section: Gcnsmentioning

confidence: 99%

GCN-Denoiser: Mesh Denoising with Graph Convolutional Networks

Shen

et al. 2022

ACM Trans. Graph.

View full text Add to dashboard Cite

In this article, we present GCN-Denoiser, a novel feature-preserving mesh denoising method based on graph convolutional networks ( GCNs ). Unlike previous learning-based mesh denoising methods that exploit handcrafted or voxel-based representations for feature learning, our method explores the structure of a triangular mesh itself and introduces a graph representation followed by graph convolution operations in the dual space of triangles. We show such a graph representation naturally captures the geometry features while being lightweight for both training and inference. To facilitate effective feature learning, our network exploits both static and dynamic edge convolutions, which allow us to learn information from both the explicit mesh structure and potential implicit relations among unconnected neighbors. To better approximate an unknown noise function, we introduce a cascaded optimization paradigm to progressively regress the noise-free facet normals with multiple GCNs. GCN-Denoiser achieves the new state-of-the-art results in multiple noise datasets, including CAD models often containing sharp features and raw scan models with real noise captured from different devices. We also create a new dataset called PrintData containing 20 real scans with their corresponding ground-truth meshes for the research community. Our code and data are available at https://github.com/Jhonve/GCN-Denoiser.

show abstract

Section: Gcnsmentioning

confidence: 99%

GCN-Denoiser: Mesh Denoising with Graph Convolutional Networks

Shen

et al. 2022

ACM Trans. Graph.

View full text Add to dashboard Cite

show abstract

“…These and other methods must also define a means of representing localized convolution kernels. Many choices are available, including localized spectral filters [Boscaini et al 2015b], B-splines [Fey et al 2018], Zernike polynomials [Sun et al 2020], wavelets [Schonsheck et al 2018], and extrinsic Euclidean convolution [Schult et al 2020].…”

Section: Neural Network On Meshesmentioning

confidence: 99%

HodgeNet

Smirnov

Solomon

2021

ACM Trans. Graph.

View full text Add to dashboard Cite

Fig. 1. Mesh segmentation results on the full-resolution MIT animation dataset. Each mesh in the dataset contains 20,000 faces (10,000 vertices). We show an example ground truth segmentation in the bottom-left. In contrast to previous works, which downsample each mesh by more than 10×, we efficiently process dense meshes both at train and test time.Constrained by the limitations of learning toolkits engineered for other applications, such as those in image processing, many mesh-based learning algorithms employ data flows that would be atypical from the perspective of conventional geometry processing. As an alternative, we present a technique for learning from meshes built from standard geometry processing modules and operations. We show that low-order eigenvalue/eigenvector computation from operators parameterized using discrete exterior calculus is amenable to efficient approximate backpropagation, yielding spectral per-element or per-mesh features with similar formulas to classical descriptors like the heat/wave kernel signatures. Our model uses few parameters, generalizes to high-resolution meshes, and exhibits performance and time complexity on par with past work.

show abstract

“…TextureNet [36] parameterizes the room surface into local planar patches in the 4-RoSy field such that standard CNNs [7] can be applied to extract high-resolution texture information from mesh facets. Schult et al [15] applied the spatial graph convolutions of dynamic filters [31], [65], [68], [79] to the union of neighborhoods in both geodesic and Euclidean domains for vertex-wise feature learning. VMNet [80] combines the SparseConvNet [56] with graph convolutional networks to learn merged features from point clouds and meshes.…”

Section: Convolution On 3d Meshesmentioning

confidence: 99%

“…Considering the impressive performance of convolutional feature learning on homogeneous grid data, i.e. images and videos [6], [7], [8], [9], [10], [11], [12], [13], researchers are also seeking alternative convolutional neural networks for 3D-mesh feature learning [1], [14], [15]. Currently this is a major research topic in geometric deep learning, which focuses on feature encoding of generic heterogeneous data in non-Euclidean space [16].…”

Section: Introductionmentioning

confidence: 99%

“…The absence of efficient modular operations for mesh processing in the modern libraries such as Tensorflow [34], Pytorch [35], requires substantial effort for deep learning application to largescale mesh inputs. This is a currently major hindrance in effective neural modeling of scene surfaces [15], [36]. Point cloud methods are unable to directly benefit from the preexisting neighborhood information in data.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Geometric Feature Learning for 3D Meshes

Lei¹,

Akhtar²,

Shah³

et al. 2021

Preprint

View full text Add to dashboard Cite

Geometric feature learning for 3D meshes is central to computer graphics and highly important for numerous vision applications. However, deep learning currently lags in hierarchical modeling of heterogeneous 3D meshes due to the lack of required operations and/or their efficient implementations. In this paper, we propose a series of modular operations for effective geometric deep learning over heterogeneous 3D meshes. These operations include mesh convolutions, (un)pooling and efficient mesh decimation. We provide open source implementation of these operations, collectively termed Picasso. The mesh decimation module of Picasso is GPU-accelerated, which can process a batch of meshes on-the-fly for deep learning. Our (un)pooling operations compute features for newly-created neurons across network layers of varying resolution. Our mesh convolutions include facet2vertex, vertex2facet, and facet2facet convolutions that exploit vMF mixture and Barycentric interpolation to incorporate fuzzy modelling. Leveraging the modular operations of Picasso, we contribute a novel hierarchical neural network, PicassoNet-II, to learn highly discriminative features from 3D meshes. PicassoNet-II accepts primitive geometrics and fine textures of mesh facets as input features, while processing full scene meshes. Our network achieves highly competitive performance for shape analysis and scene parsing on a variety of benchmarks. We release Picasso and PicassoNet-II on Github.

show abstract

DualConvMesh-Net: Joint Geodesic and Euclidean Convolutions on 3D Meshes

Cited by 86 publications

References 30 publications

GCN-Denoiser: Mesh Denoising with Graph Convolutional Networks

GCN-Denoiser: Mesh Denoising with Graph Convolutional Networks

HodgeNet

Geometric Feature Learning for 3D Meshes

Contact Info

Product

Resources

About