Selected work

Representative Papers

Completely Independent Spanning Trees in $k$-Outerplanar Triangulated Discs

Jun 10, 2026

This study investigates the existence of two completely independent spanning trees (CISTs) in $k$-outerplanar triangulations. Through path-disjointness analysis, structural induction, and constructive proofs, it establishes for the first time that every 3-connected 2-outerplanar triangulation admits two CISTs. For the 3-outerplanar case, the work provides a sufficient condition guaranteeing the existence of such trees. Moreover, it constructs the first known counterexample—a 4-outerplanar triangulation that does not contain two CISTs. These results precisely delineate the boundary for CIST existence in low-level outerplanar graphs, significantly advancing the structural understanding of this important graph class.

0 citationsRead paper

AtteConDA: Attention-Based Conflict Suppression in Multi-Condition Diffusion Models and Synthetic Data Augmentation

May 10, 2026

This work addresses the challenge of structural distortion in multi-condition image generation caused by conflicting conditioning signals, which limits the utility of synthetic data for high-level autonomous driving tasks. To mitigate this issue, the authors propose an attention-based multi-condition diffusion model that explicitly suppresses condition conflicts during generation by jointly leveraging semantic segmentation maps, depth maps, and edge information. The method achieves, for the first time, high-fidelity structural preservation in multi-condition image synthesis within autonomous driving scenarios and introduces a dedicated evaluation framework. Experimental results demonstrate that the generated images significantly outperform existing approaches in structural consistency, effectively alleviating data scarcity and establishing a new benchmark for future research.

0 citationsRead paper

An Upper Bound for the Double Domination Number in Maximal Outerplanar Graphs

Mar 03, 2026

This study addresses the upper bound of the double domination number—defined as the minimum size of a set in which every vertex is dominated by at least two vertices in the set—for maximal outerplanar graphs. By integrating domination theory with structural properties specific to maximal outerplanar graphs, particularly the number $k$ of pairs of consecutive vertices on the outer face whose distance is at least three, the authors establish the tight upper bound $\gamma_{\times 2}(G) \leq (n + k)/2$. This work provides the first complete and rigorous proof of this previously unverified bound, thereby filling a notable gap in the theoretical literature and advancing the understanding of domination properties in maximal outerplanar graphs.

0 citationsRead paper

Motor Imagery Classification Using Feature Fusion of Spatially Weighted Electroencephalography

Nov 13, 2025

To address the low classification accuracy and high computational complexity in motor imagery (MI)-based brain–computer interfaces (BCIs) using multi-channel electroencephalography (EEG), this paper proposes a novel channel selection and multi-domain feature fusion framework grounded in functional brain regional characteristics. First, electrodes are grouped according to anatomically and functionally defined brain regions, and key channels are selected based on inter-regional correlation analysis. Subsequently, discriminative features are extracted from three complementary domains: spatial features via common spatial patterns (CSP), clustering-based features using fuzzy C-means, and manifold features derived from tangent space mapping on Riemannian manifolds. This fusion strategy enhances feature discriminability while mitigating redundancy. Finally, a support vector machine (SVM) classifier is employed. Evaluated on BCI Competition IV datasets 2a and 2b, the method achieves average classification accuracies of 90.77% and 84.50%, respectively—outperforming state-of-the-art approaches.

0 citationsRead paper

Euclidean k-center Fair Clusterings

Nov 10, 2025

This paper studies the fair $k$-clustering problem for a colored point set $P$ in the plane: partition $P$ into $k$ clusters such that, for each color $q$, the number of points of color $q$ in every cluster lies strictly between given lower and upper bounds $l(q)$ and $u(q)$, ensuring balanced representation of demographic groups (e.g., gender, ethnicity) within each cluster. We introduce and formally define the “fair $k$-center problem under lower- and upper-bound constraints,” the first formulation to enforce strict per-cluster fairness via hard bounds—contrasting prior work relying on global proportionality or soft constraints. Leveraging geometric partitioning and integer programming relaxation, we design the first polynomial-time approximation algorithm that guarantees feasibility while minimizing the maximum cluster radius. We prove a constant-factor approximation ratio and validate the algorithm’s effectiveness and practicality through empirical evaluation.

0 citationsRead paper

Recent publications

Latest Papers

Completely Independent Spanning Trees in $k$-Outerplanar Triangulated Discs

Jun 10, 2026

This study investigates the existence of two completely independent spanning trees (CISTs) in $k$-outerplanar triangulations. Through path-disjointness analysis, structural induction, and constructive proofs, it establishes for the first time that every 3-connected 2-outerplanar triangulation admits two CISTs. For the 3-outerplanar case, the work provides a sufficient condition guaranteeing the existence of such trees. Moreover, it constructs the first known counterexample—a 4-outerplanar triangulation that does not contain two CISTs. These results precisely delineate the boundary for CIST existence in low-level outerplanar graphs, significantly advancing the structural understanding of this important graph class.

0 citationsRead paper

AtteConDA: Attention-Based Conflict Suppression in Multi-Condition Diffusion Models and Synthetic Data Augmentation

May 10, 2026

This work addresses the challenge of structural distortion in multi-condition image generation caused by conflicting conditioning signals, which limits the utility of synthetic data for high-level autonomous driving tasks. To mitigate this issue, the authors propose an attention-based multi-condition diffusion model that explicitly suppresses condition conflicts during generation by jointly leveraging semantic segmentation maps, depth maps, and edge information. The method achieves, for the first time, high-fidelity structural preservation in multi-condition image synthesis within autonomous driving scenarios and introduces a dedicated evaluation framework. Experimental results demonstrate that the generated images significantly outperform existing approaches in structural consistency, effectively alleviating data scarcity and establishing a new benchmark for future research.

0 citationsRead paper

An Upper Bound for the Double Domination Number in Maximal Outerplanar Graphs

Mar 03, 2026

This study addresses the upper bound of the double domination number—defined as the minimum size of a set in which every vertex is dominated by at least two vertices in the set—for maximal outerplanar graphs. By integrating domination theory with structural properties specific to maximal outerplanar graphs, particularly the number $k$ of pairs of consecutive vertices on the outer face whose distance is at least three, the authors establish the tight upper bound $\gamma_{\times 2}(G) \leq (n + k)/2$. This work provides the first complete and rigorous proof of this previously unverified bound, thereby filling a notable gap in the theoretical literature and advancing the understanding of domination properties in maximal outerplanar graphs.

0 citationsRead paper

Motor Imagery Classification Using Feature Fusion of Spatially Weighted Electroencephalography

Nov 13, 2025

To address the low classification accuracy and high computational complexity in motor imagery (MI)-based brain–computer interfaces (BCIs) using multi-channel electroencephalography (EEG), this paper proposes a novel channel selection and multi-domain feature fusion framework grounded in functional brain regional characteristics. First, electrodes are grouped according to anatomically and functionally defined brain regions, and key channels are selected based on inter-regional correlation analysis. Subsequently, discriminative features are extracted from three complementary domains: spatial features via common spatial patterns (CSP), clustering-based features using fuzzy C-means, and manifold features derived from tangent space mapping on Riemannian manifolds. This fusion strategy enhances feature discriminability while mitigating redundancy. Finally, a support vector machine (SVM) classifier is employed. Evaluated on BCI Competition IV datasets 2a and 2b, the method achieves average classification accuracies of 90.77% and 84.50%, respectively—outperforming state-of-the-art approaches.

0 citationsRead paper

Euclidean k-center Fair Clusterings

Nov 10, 2025

This paper studies the fair $k$-clustering problem for a colored point set $P$ in the plane: partition $P$ into $k$ clusters such that, for each color $q$, the number of points of color $q$ in every cluster lies strictly between given lower and upper bounds $l(q)$ and $u(q)$, ensuring balanced representation of demographic groups (e.g., gender, ethnicity) within each cluster. We introduce and formally define the “fair $k$-center problem under lower- and upper-bound constraints,” the first formulation to enforce strict per-cluster fairness via hard bounds—contrasting prior work relying on global proportionality or soft constraints. Leveraging geometric partitioning and integer programming relaxation, we design the first polynomial-time approximation algorithm that guarantees feasibility while minimizing the maximum cluster radius. We prove a constant-factor approximation ratio and validate the algorithm’s effectiveness and practicality through empirical evaluation.

0 citationsRead paper