Simultaneous Low-rank Component and Graph Estimation for High-dimensional Graph Signals: Application to Brain Imaging
This work addresses a domain-specific problem in brain imaging analysis, offering an incremental improvement by simultaneously learning low-rank components and refining graphs for noisy datasets.
The authors tackled the problem of estimating the intrinsic low-rank component and underlying graph from high-dimensional, graph-smooth, and grossly-corrupted data when the graph is unknown, achieving encouraging performance in unsupervised and supervised classification tasks on synthetic and real brain imaging data.
We propose an algorithm to uncover the intrinsic low-rank component of a high-dimensional, graph-smooth and grossly-corrupted dataset, under the situations that the underlying graph is unknown. Based on a model with a low-rank component plus a sparse perturbation, and an initial graph estimation, our proposed algorithm simultaneously learns the low-rank component and refines the graph. Our evaluations using synthetic and real brain imaging data in unsupervised and supervised classification tasks demonstrate encouraging performance.