fine-grained thin-structure localization (wire/filament detection)

Existing superpixel methods struggle to simultaneously achieve computational efficiency, boundary adherence, and superpixel regularity. To address this, we propose SCALP—a novel superpixel segmentation framework that introduces, for the first time, a linear path distance metric to replace conventional Euclidean distance within an iterative clustering scheme. This metric models the shortest contour-aware path from each pixel to its cluster centroid, thereby jointly optimizing compactness, regularity, and boundary conformity through geometric and semantic boundary co-constraints. Evaluated on the BSD dataset, SCALP outperforms all state-of-the-art methods: it achieves superior superpixel quality (measured by Undersegmentation Error and Boundary Recall) and state-of-the-art contour detection performance (ODS and OIS scores), while maintaining high computational efficiency. Our key contribution lies in breaking the inherent trade-off among compactness, regularity, and boundary adherence—establishing a new metric paradigm and efficient implementation for superpixel segmentation.

This work addresses the fragmentation problem in segmenting fine structures—such as wires, cracks, and lane markings—caused by traditional pixel-wise representations that disrupt topological connectivity. To preserve structural continuity, the authors propose a topology-preserving, parameterized graph minor representation that compresses the input image into super-nodes via a boundary-aligned contraction criterion. This approach achieves substantial dimensionality reduction while rigorously maintaining the connectivity of fine structures. A lightweight graph neural network is then employed for classification, followed by bidirectional pixel-to-graph mapping to enable full-resolution inference. Evaluated on TTPLA, CrackSeg9k, and SkyScapes Lane datasets, the method matches or surpasses state-of-the-art domain-specific approaches in Dice, IoU, and Boundary F1 scores, while reducing mask fragmentation by at least 4.6×.

This study addresses the challenge of automatically distinguishing individual ply instances in high-resolution microscopic images of carbon fiber–reinforced polymers (CFRP), where semantic boundaries between adjacent plies are often ambiguous. To this end, the authors propose a global information–driven shortest-path approach that transforms a semantic segmentation mask into a precise ply-level instance segmentation. By modeling the image as a graph and leveraging shortest-path algorithms, the method accurately delineates ply boundaries and assigns each fiber pixel to its corresponding ply—without requiring additional annotations. This enables quantitative analysis of local fiber volume fractions, ply thicknesses, and interlayer spacing. The approach demonstrates robust performance on CFRP images containing artificial gaps, complex stacking sequences, and through-thickness cracks, effectively visualizing and quantifying manufacturing-induced heterogeneities to support process optimization and structure–property correlation studies.

This study addresses the challenge of detecting subtle, low-contrast defects embedded within dense backgrounds in printed circuit board (PCB) inspection. To this end, the authors propose a two-stage detection framework: first, a structure-guided hybrid masked sparse convolution is employed for pretraining to learn structural priors inherent to PCB layouts; second, during fine-tuning, a spatial continuity regularization is introduced to enhance the compactness and coherence of predictions for elongated defect regions. By innovatively integrating structural prior modeling with spatial constraints, the method achieves state-of-the-art performance on the DsPCBSD+ dataset, attaining 85.5% mAP₀.₅ and 52.3% mAP₀.₅:₀.₉₅, significantly outperforming existing strong baselines.

Existing Subspace Constrained Mean Shift (SCMS) algorithms for filamentary structure extraction in point clouds lack global convergence guarantees and often stagnate at local optima. Method: This paper models filaments as ridges of the underlying density function and proposes two novel algorithms with rigorous theoretical convergence guarantees: (i) a ridge detection method leveraging gradient and Hessian geometric features, and (ii) a convergence framework integrating iterative projection optimization with Lyapunov stability analysis. Contribution/Results: We provide the first globally convergent proof for ridge estimation under subspace constraints, overcoming SCMS’s robustness limitations. Extensive experiments on synthetic and real-world point cloud data demonstrate substantial improvements in convergence stability and ridge-line localization accuracy. Theoretically, we prove that the generated iteration sequence converges almost surely to the true ridge set.

本文针对血管图像中图匹配不一致的问题，提出了一种先匹配后联合精炼的策略，以提高纵向血管图像的匹配精度和一致性。

This work addresses the challenge of precise sensor calibration required for large-scale robotic perception by proposing a high-coverage, calibration-free obstacle detection method. The approach leverages computer-aided design (CAD) to procedurally generate “skin units” that conform to the robot’s non-developable surfaces, enabling the integration of arbitrarily sized circuit boards to fix sensor positions with known spatial coordinates. By combining CAD, 3D-printed structural elements, and time-of-flight (ToF) imaging arrays, the method successfully constructs near-field obstacle point clouds on a Franka Research 3 robotic arm. This enables extensive, efficient perception without prior calibration, significantly streamlining deployment and enhancing system practicality.

This work addresses the challenge of topological errors—such as spurious connections and centerline fractures—in automatic extraction of 3D tubular structure skeletons, which commonly arise from noise or missing data and are inefficient and error-prone to correct manually. The authors propose a lightweight semi-automatic correction method: given user-specified start and end points, it performs locally stable propagation via component-wise minimum spanning trees and bridges gaps or resolves ambiguous connections using filtered 3D Delaunay edge graphs. Candidate paths are ranked through a scoring mechanism that integrates directional continuity, spatial proximity, component consistency, and goal-directedness. Implemented in C++ with Libigl, the interactive system effectively repairs typical artifacts like “crossings” and “breaks” in cerebral vasculature data, producing ordered polylines suitable for downstream processing and demonstrating both practical utility and robustness.

This work addresses the challenge of efficiently rendering sparse microstructured materials—such as fibrous or brushed metals—in volume rendering, which typically demand high-resolution representations yet incur substantial computational costs with conventional voxelization and multiscale rendering approaches. To overcome this, the authors propose an efficient parallel voxelization method that integrates hierarchical SGGX clustering to construct a level-of-detail (LoD) representation, significantly accelerating multiscale data aggregation and rendering. Implemented in CUDA, the approach supports both triangle meshes and explicit fiber models and is embedded within an SGGX distribution–driven LoD path tracing framework. Experimental results demonstrate that the method achieves a superior balance between rendering quality and performance across a range of microstructured materials compared to baseline techniques.

This study addresses the fragmented conceptualization and inconsistent implementation standards of existing watershed segmentation algorithms, which hinder research reproducibility and application efficiency. To overcome these limitations, this work presents the first compact integration of diverse watershed theories and algorithms. By leveraging weighted graphs, Kruskal-based minimum spanning trees, and connected component analysis, it constructs an end-to-end segmentation pipeline encompassing both supervised and unsupervised paradigms alongside multiple seed computation variants. The primary contribution lies in establishing a clear, unified implementation standard that serves as a highly reproducible framework guide for watershed segmentation. Ultimately, this systematic consolidation significantly enhances both the theoretical understanding and the engineering implementation efficiency of watershed algorithms within the field.