arXiv subjects

Kuan-Wei Chen

Publications and source records attributed to Kuan-Wei Chen.

2 recordsLinked to original sources

arXiv2026

Global Complete Synchronization in Networks of Identical Stuart--Landau Oscillators

We study global complete synchronization in finite networks of diffusively coupled Stuart--Landau oscillators with identical natural frequencies. The underlying graph is arbitrary, connected, and undirected. In contrast with phase-only models, the amplitudes evolve dynamically and may approach zero, where the polar phase equations become singular. We derive explicit sufficient conditions that keep every amplitude uniformly separated from zero and then show that the classical invariant-arc mechanism for identical Kuramoto oscillators survives in this amplitude-inclusive setting. Two synchronization criteria are obtained, combining amplitude bounds, phase confinement, an energy identity, and the asymptotic structure of the synchronized manifold. Under either criterion, every oscillator converges exponentially to the same periodic orbit of radius $\sqrtμ$ and common frequency. The result provides a graph-level extension of the invariant-arc synchronization mechanism from identical Kuramoto networks to Stuart--Landau networks while retaining explicit control of the amplitude dynamics.

math.DS↗

arXiv2023

Seg2Reg: Differentiable 2D Segmentation to 1D Regression Rendering for 360 Room Layout Reconstruction

State-of-the-art single-view 360-degree room layout reconstruction methods formulate the problem as a high-level 1D (per-column) regression task. On the other hand, traditional low-level 2D layout segmentation is simpler to learn and can represent occluded regions, but it requires complex post-processing for the targeting layout polygon and sacrifices accuracy. We present Seg2Reg to render 1D layout depth regression from the 2D segmentation map in a differentiable and occlusion-aware way, marrying the merits of both sides. Specifically, our model predicts floor-plan density for the input equirectangular 360-degree image. Formulating the 2D layout representation as a density field enables us to employ `flattened' volume rendering to form 1D layout depth regression. In addition, we propose a novel 3D warping augmentation on layout to improve generalization. Finally, we re-implement recent room layout reconstruction methods into our codebase for benchmarking and explore modern backbones and training techniques to serve as the strong baseline. Our model significantly outperforms previous arts. The code will be made available upon publication.

cs.CV↗