Wei Wu

Handwritten identification marks on steel plate end faces are important for material traceability, inventory checking, and production scheduling in large steel storage yards. Their localization remains challenging because field images often contain uneven illumination, specular reflection, rust, oil contamination, densely adjacent targets, and arbitrary target orientations. This study presents an application-oriented pipeline for oriented localization of handwritten marks on densely stacked steel plates. The pipeline combines task-oriented image enhancement, high-resolution sliding-window inference, and a YOLOv11 oriented bounding-box detector (YOLOv11-OBB) adapted with attention enhancement, adjusted multi-scale fusion, and Scylla-IoU (SIoU)-based box regression. A real steel-yard dataset was annotated with oriented bounding boxes and split at the source-image level before patch extraction. The held-out evaluation subset was used both for checkpoint selection and final performance reporting under the same protocol. On this dataset, the final detector achieved 89.9% precision, 88.3% recall, 91.8% mAP@0.5, and 72.0% mAP@0.5:0.95. These results support the practical value of oriented localization and high-resolution inference for weak handwritten regions in complex industrial scenes. Broader detector benchmarking and downstream recognition performance remain to be validated in future work.

Aiming at the core bottleneck of the traditional numerical manifold method (NMM)—difficulty in 3D extension caused by its reliance on complex geometric Boolean operations when solving 3D continuous‐discontinuous problems—this paper proposes an element‐free numerical manifold method (E‐NMM) that does not generate manifold element (ME). The method replaces the ME generation process with voxel discretization technology, and uses numerical manifold point (NMP) to undertake integral representation and post‐processing functions, completely avoiding 3D Boolean operations; it achieves accurate reconstruction of physical cover (PC) by combining enhanced connected component analysis. The effectiveness of E‐NMM and its applicability to complex engineering problems are verified through multiple sets of numerical examples. E‐NMM retains the core feature of NMM, that is, the unified solution for continuous and discontinuous problems, and provides an efficient and reliable numerical solution tool for 3D complex continuous‐discontinuous problems in fields such as rock mechanics and tunnel engineering.

Dynamic scene reconstruction has achieved significant milestones with the advent of 3D Gaussian Splatting (3DGS). However, extending this technology from geometric reconstruction to semantic understanding in dynamic environments remains a challenge. Existing methods often rely on external 2D trackers, which lead to temporal inconsistencies and semantic drift, or suffer from the high computational costs of high-dimensional feature fields. In this paper, we propose a novel framework, Gaussian Semantic Segmentation based on Color and Shape Deformation Fields (GSSBC), to address these issues. Building upon our GBC dynamic scene representation, we bind learnable semantic features to deformable Gaussian primitives. We introduce a spatiotemporal contrastive learning strategy guided by the Segment Anything Model (SAM) to enforce semantic consistency without explicit tracking. Furthermore, we employ a density-based clustering algorithm with label propagation to extract discrete object entities efficiently. Experimental results on the HyperNeRF and Neu3D datasets demonstrate that our method achieves superior segmentation accuracy and spatiotemporal stability compared to state-of-the-art approaches, enabling effective semantic understanding in complex dynamic scenes.

In the area of vehicle data handling in cloud storage systems, the analyzing and enhancing of vehicle traits efficiently are at the heart of the optimization of both storage and retrieval processes. We introduce a Dense Pattern Fusion (DPF), a new set of tools for large-scale vehicle data analysis through hypergraph learning methods with an adaptive vehicle feature fusion technique. The suggested system divides and assesses vehicle patterns in a procedure by finding groups of cars sharing the same operational habits and environmental characteristics. Moreover, it skillfully handles the challenges that usually come with the variated nature of vehicle data coming from different districts and comprising various attributes. The framework utilizes hypergraph-based feature fusion to detect the quality vehicle traits that mirror their specific operational patterns and environmental effects. The use of a probabilistic model makes it possible for Dense Pattern Fusion to place the attributes of each vehicle in a joint latent space, thus providing a more comprehensive view of its operation movements. The underlying system develops a large and detailed graph showing the interrelations and likenesses among the different types of cars, revealing thick subgraphs or “clusters” of cars sharing common characteristics. Then these clusters are analyzed for insights, which results in the data analysis being more efficient. The adaptive merging of car properties improves the entire data handling process, thus the power of cloud storage systems is better utilized. The tests carried out on a rich literacy dataset consisting of car information from different places prove the system’s effectiveness in refining car data analysis. The findings highlight Dense Pattern Fusion’s capacity to make car data management coincide with the needs of cloud storage, showing its ability to scale and adapt when working with large, heterogeneous datasets of cars. This novel technique provides a solid answer to the problem of cloud vehicle data analysis, leading to less expensive storage and better data-access strategies. The proposed Dense Pattern Fusion model has shown further measurable improvements in cloud storage efficiency by eliminating unnecessary vehicle data, speeding up retrieval time and enabling a scalable data organization across distributed storage infrastructures that are supported.