Manifold Matching using Shortest-Path Distance and Joint Neighborhood Selection
arXiv:1412.4098 · doi:10.1016/j.patrec.2017.04.005
Abstract
Matching datasets of multiple modalities has become an important task in data analysis. Existing methods often rely on the embedding and transformation of each single modality without utilizing any correspondence information, which often results in sub-optimal matching performance. In this paper, we propose a nonlinear manifold matching algorithm using shortest-path distance and joint neighborhood selection. Specifically, a joint nearest-neighbor graph is built for all modalities. Then the shortest-path distance within each modality is calculated from the joint neighborhood graph, followed by embedding into and matching in a common low-dimensional Euclidean space. Compared to existing algorithms, our approach exhibits superior performance for matching disparate datasets of multiple modalities.
arXiv comment13 pages, 8 figures, 2 tables
References in corpus (8)
- Kernel methods in machine learning
- Brownian distance covariance
- Manifold Learning: The Price of Normalization
- Seeded graph matching for correlated Erdős-Rényi graphs
- MIGRAINE: MRI Graph Reliability Analysis and Inference for Connectomics
- Generalized Canonical Correlation Analysis for Classification
- Robust Multimodal Graph Matching: Sparse Coding Meets Graph Matching
- On the Incommensurability Phenomenon
Cited by in corpus (5)
- Discovering and Deciphering Relationships Across Disparate Data Modalities
- Synergistic Graph Fusion via Encoder Embedding
- Encoder Embedding for General Graph and Node Classification
- Discovering the Signal Subgraph: An Iterative Screening Approach on Graphs
- Network Dependence Testing via Diffusion Maps and Distance-Based Correlations