MeshWalker: Deep Mesh Understanding by Random Walks

Lahav, Alon; Tal, Ayellet

doi:10.48550/arxiv.2006.05353

Cited by 1 publication

(1 citation statement)

References 66 publications

(81 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Other methods include [de Haan et al 2020], which proposes anisotropic gauge-invariant kernels using a message passing scheme built from parallel transport; [Lahav and Tal 2020], an RNN-based approach employing random walks; [Schneider et al 2020], which improves MeshCNN's memory efficiency and resilience to class imbalance for medical applications; and [Budninskiy et al 2020], which optimizes for a graph Laplacian parameterized by edge features.…”

Section: Neural Network On Meshesmentioning

confidence: 99%

HodgeNet: Learning Spectral Geometry on Triangle Meshes

Smirnov,

Solomon

2021

Preprint

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. CCS Concepts: • Computing methodologies → Shape analysis; Mesh geometry models; Neural networks.

show abstract

Section: Neural Network On Meshesmentioning

confidence: 99%