ResearchTrend.AI
  • Papers
  • Communities
  • Events
  • Blog
  • Pricing
Papers
Communities
Social Events
Terms and Conditions
Pricing
Parameter LabParameter LabTwitterGitHubLinkedInBlueskyYoutube

© 2025 ResearchTrend.AI, All rights reserved.

  1. Home
  2. Papers
  3. 1907.09138
53
89
v1v2 (latest)

Feature Graph Learning for 3D Point Cloud Denoising

22 July 2019
Wei Hu
Xiang Gao
Gene Cheung
Zongming Guo
ArXiv (abs)PDFHTML
Abstract

Identifying an appropriate underlying graph kernel that reflects pairwise similarities is critical in many recent graph spectral signal restoration schemes, including image denoising, dequantization, and contrast enhancement. Existing graph learning algorithms compute the most likely entries of a properly defined graph Laplacian matrix L\mathbf{L}L, but require a large number of signal observations z\mathbf{z}z's for a stable estimate. In this work, we assume instead the availability of a relevant feature vector fi\mathbf{f}_ifi​ per node iii, from which we compute an optimal feature graph via optimization of a feature metric. Specifically, we alternately optimize the diagonal and off-diagonal entries of a Mahalanobis distance matrix M\mathbf{M}M by minimizing the graph Laplacian regularizer (GLR) z⊤Lz\mathbf{z}^{\top} \mathbf{L} \mathbf{z}z⊤Lz, where edge weight is wi,j=exp⁡{−(fi−fj)⊤M(fi−fj)}w_{i,j} = \exp\{-(\mathbf{f}_i - \mathbf{f}_j)^{\top} \mathbf{M} (\mathbf{f}_i - \mathbf{f}_j) \}wi,j​=exp{−(fi​−fj​)⊤M(fi​−fj​)}, given a single observation z\mathbf{z}z. We optimize diagonal entries via proximal gradient (PG), where we constrain M\mathbf{M}M to be positive definite (PD) via linear inequalities derived from the Gershgorin circle theorem. To optimize off-diagonal entries, we design a block descent algorithm that iteratively optimizes one row and column of M\mathbf{M}M. To keep M\mathbf{M}M PD, we constrain the Schur complement of sub-matrix M2,2\mathbf{M}_{2,2}M2,2​ of M\mathbf{M}M to be PD when optimizing via PG. Our algorithm mitigates full eigen-decomposition of M\mathbf{M}M, thus ensuring fast computation speed even when feature vector fi\mathbf{f}_ifi​ has high dimension. To validate its usefulness, we apply our feature graph learning algorithm to the problem of 3D point cloud denoising, resulting in state-of-the-art performance compared to competing schemes in extensive experiments.

View on arXiv
Comments on this paper