Hao Shu

Hao Shu

Image edge detection (ED) requires specialized architectures, reliable supervision, and rigorous evaluation criteria to ensure accurate localization. In this work, we present a framework for high-precision ED that jointly addresses architectural design, data supervision, and evaluation consistency. We propose SDPED, a compact ED model built upon Cascaded Skipping Density Blocks (CSDB), motivated by a task-adaptive architectural transfer from image super-resolution. By re-engineering texture-oriented structures for ED, SDPED effectively differentiates textures from edges while preserving fine spatial precision. Extensive experiments on four benchmark datasets (BRIND, UDED, MDBD, and BIPED2) demonstrate consistent performance improvements, particularly in Average Precision (AP), with gains of up to 22.5% on MDBD and 11.8% on BIPED2. In addition, we introduce an ideal-prior guidance strategy that incorporates noiseless data into training by treating labels as noise-free samples, providing a practical means to mitigate the subjectivity and noise inherent in human annotations. To enable fair and resolution-independent evaluation, we further adopt a fixed-pixel criterion for assessing localization accuracy. Overall, this work offers a coherent solution for high-precision ED and provides insights applicable to precision-oriented modeling in low-level and soft-computing-based vision tasks. Codes can be found on https://github.com/Hao-B-Shu/SDPED.

Hao Shu, Jicheng Li, Yu Jin et al.

Non-random missing data is a ubiquitous yet undertreated flaw in multidimensional time series, fundamentally threatening the reliability of data-driven analysis and decision-making. Pure low-rank tensor completion, as a classical data recovery method, falls short in handling non-random missingness, both methodologically and theoretically. Hankel-structured tensor completion models provide a feasible approach for recovering multidimensional time series with non-random missing patterns. However, most Hankel-based multidimensional data recovery methods both suffer from unclear sources of Hankel tensor low-rankness and lack an exact recovery theory for non-random missing data. To address these issues, we propose the temporal isometric delay-embedding transform, which constructs a Hankel tensor whose low-rankness is naturally induced by the smoothness and periodicity of the underlying time series. Leveraging this property, we develop the \textit{Low-Rank Tensor Completion with Temporal Isometric Delay-embedding Transform} (LRTC-TIDT) model, which characterizes the low-rank structure under the \textit{Tensor Singular Value Decomposition} (t-SVD) framework. Once the prescribed non-random sampling conditions and mild incoherence assumptions are satisfied, the proposed LRTC-TIDT model achieves exact recovery, as confirmed by simulation experiments under various non-random missing patterns. Furthermore, LRTC-TIDT consistently outperforms existing tensor-based methods across multiple real-world tasks, including network flow reconstruction, urban traffic estimation, and temperature field prediction. Our implementation is publicly available at https://github.com/HaoShu2000/LRTC-TIDT.

Hao Shu

This work introduces a learning-enhanced observer (LEO) for linear time-invariant systems with uncertain dynamics. Rather than relying solely on nominal models, the proposed framework treats the system matrices as optimizable variables and refines them through gradient-based minimization of a steady-state output discrepancy loss. The resulting data-informed surrogate model enables the construction of an improved observer that effectively compensates for moderate parameter uncertainty while preserving the structure of classical designs. Extensive Monte Carlo studies across diverse system dimensions show systematic and statistically significant reductions, typically exceeding 15\%, in normalized estimation error for both open-loop and Luenberger observers. These results demonstrate that modern learning mechanisms can serve as a powerful complement to traditional observer design, yielding more accurate and robust state estimation in uncertain systems. Codes are available at https://github.com/Hao-B-Shu/LTI_LEO.

Hao Shu, Jicheng Li, Yu Jin et al.

In recent years, the prediction of multidimensional time series data has become increasingly important due to its wide-ranging applications. Tensor-based prediction methods have gained attention for their ability to preserve the inherent structure of such data. However, existing approaches, such as tensor autoregression and tensor decomposition, often have consistently failed to provide clear assertions regarding the number of samples that can be exactly predicted. While matrix-based methods using nuclear norms address this limitation, their reliance on matrices limits accuracy and increases computational costs when handling multidimensional data. To overcome these challenges, we reformulate multidimensional time series prediction as a deterministic tensor completion problem and propose a novel theoretical framework. Specifically, we develop a deterministic tensor completion theory and introduce the Temporal Convolutional Tensor Nuclear Norm (TCTNN) model. By convolving the multidimensional time series along the temporal dimension and applying the tensor nuclear norm, our approach identifies the maximum forecast horizon for exact predictions. Additionally, TCTNN achieves superior performance in prediction accuracy and computational efficiency compared to existing methods across diverse real-world datasets, including climate temperature, network flow, and traffic ride data. Our implementation is publicly available at https://github.com/HaoShu2000/TCTNN.

Hao Shu

Density-based clustering is the most popular clustering algorithm since it can identify clusters of arbitrary shape as long as they are separated by low-density regions. However, a high-density region that is not separated by low-density ones might also have different structures belonging to multiple clusters. As far as we know, all previous density-based clustering algorithms fail to detect such structures. In this paper, we provide a novel density-based clustering scheme to address this problem. It is the rst clustering algorithm that can detect meticulous structures in a high-density region that is not separated by low-density ones and thus extends the range of applications of clustering. The algorithm employs secondary directed differential, hierarchy, normalized density, as well as the self-adaption coefficient, called Structure Detecting Cluster by Hierarchical Secondary Directed Differential with Normalized Density and Self-Adaption, dubbed SDC-HSDD-NDSA. Experiments on synthetic and real datasets are implemented to verify the effectiveness, robustness, and granularity independence of the algorithm, and the scheme is compared to unsupervised schemes in the Python package Scikit-learn. Results demonstrate that our algorithm outperforms previous ones in many situations, especially significantly when clusters have regular internal structures. For example, averaging over the eight noiseless synthetic datasets with structures employing ARI and NMI criteria, previous algorithms obtain scores below 0.6 and 0.7, while the presented algorithm obtains scores higher than 0.9 and 0.95, respectively.

Hao Shu

Edge detection (ED) is a fundamental component in many computer vision tasks, yet achieving both high quantitative accuracy and perceptual quality remains a significant challenge. In this paper, we propose the Symmetrization Weighted Binary Cross-Entropy (SWBCE) loss function, a novel approach that addresses this issue by leveraging the inherent asymmetry in human edge perception, where edge decisions require stronger justification than non-edge ones. By balancing label-guided and prediction-guided learning, SWBCE maintains high edge recall while effectively suppressing false positives. Extensive experiments across multiple datasets and baseline models, along with comparisons to prior loss functions, demonstrate that our method consistently improves both the quantitative metrics and perceptual quality of ED results. These findings underscore the effectiveness of SWBCE for high-quality edge prediction and its potential applicability to related vision tasks.

Hao Shu

Deep learning has significantly advanced image edge detection (ED), primarily through improved feature extraction. However, most existing ED models apply uniform feature fusion across all pixels, ignoring critical differences between regions such as edges and textures. To address this limitation, we propose the Extractor-Selector (E-S) paradigm, a novel framework that introduces pixel-wise feature selection for more adaptive and precise fusion. Unlike conventional image-level fusion that applies the same convolutional kernel to all pixels, our approach dynamically selects relevant features at each pixel, enabling more refined edge predictions. The E-S framework can be seamlessly integrated with existing ED models without architectural changes, delivering substantial performance gains. It can also be combined with enhanced feature extractors for further accuracy improvements. Extensive experiments across multiple benchmarks confirm that our method consistently outperforms baseline ED models. For instance, on the BIPED2 dataset, the proposed framework can achieve over 7$\%$ improvements in ODS and OIS, and 22$\%$ improvements in AP, demonstrating its effectiveness and superiority.

Hao Shu

Edge detection (ED) remains a fundamental task in computer vision, yet its performance is often hindered by the ambiguous nature of non-edge pixels near object boundaries. The widely adopted Weighted Binary Cross-Entropy (WBCE) loss treats all non-edge pixels uniformly, overlooking the structural nuances around edges and often resulting in blurred predictions. In this paper, we propose the Edge-Boundary-Texture (EBT) loss, a novel objective that explicitly divides pixels into three categories, edge, boundary, and texture, and assigns each a distinct supervisory weight. This tri-class formulation enables more structured learning by guiding the model to focus on both edge precision and contextual boundary localization. We theoretically show that the EBT loss generalizes the WBCE loss, with the latter becoming a limit case. Extensive experiments across multiple benchmarks demonstrate the superiority of the EBT loss both quantitatively and perceptually. Furthermore, the consistent use of unified hyperparameters across all models and datasets, along with robustness to their moderate variations, indicates that the EBT loss requires minimal fine-tuning and is easily deployable in practice.

Hao Shu, Jicheng Li, Tianyv Lei et al.

In real-world scenarios, spatiotemporal traffic data frequently experiences dual degradation from missing values and noise caused by sensor malfunctions and communication failures. Therefore, effective data recovery methods are essential to ensure the reliability of downstream data-driven applications. while classical tensor completion methods have been widely adopted, they are incapable of modeling noise, making them unsuitable for complex scenarios involving simultaneous data missingness and noise interference. Existing Robust Tensor Completion (RTC) approaches offer potential solutions by separately modeling the actual tensor data and noise. However, their effectiveness is often constrained by the over-relaxation of convex rank surrogates and the suboptimal utilization of local consistency, leading to inadequate model accuracy. To address these limitations, we first introduce the tensor L1-L2 norm, a novel non-convex tensor rank surrogate that functions as an effective low-rank representation tool. Leveraging an advanced feature fusion strategy, we further develop the gradient tensor L1-L2 norm by incorporating the tensor L1-L2 norm in the gradient domain. By integrating the gradient tensor nuclear L1-L2 norm into the RTC framework, we propose the Robust Tensor Completion via Gradient Tensor Nuclear L1-L2 Norm (RTC-GTNLN) model, which not only fully exploits both global low-rankness and local consistency without trade-off parameter, but also effectively handles the dual degradation challenges of missing data and noise in traffic data. Extensive experiments conducted on multiple real-world traffic datasets demonstrate that the RTC-GTNLN model consistently outperforms existing state-of-the-art methods in complex recovery scenarios involving simultaneous missing values and noise.

Hao Shu

Image edge detection (ED) faces a fundamental mismatch between training and inference: models are trained using continuous-valued outputs but evaluated using binary predictions. This misalignment, caused by the non-differentiability of binarization, weakens the link between learning objectives and actual task performance. In this paper, we propose a theoretical method to design a Binarization-Aware Adjuster (BAA), which explicitly incorporates binarization behavior into gradient-based optimization. At the core of BAA is a novel loss adjustment mechanism based on a Distance Weight Function (DWF), which reweights pixel-wise contributions according to their correctness and proximity to the decision boundary. This emphasizes decision-critical regions while down-weighting less influential ones. We also introduce a self-adaptive procedure to estimate the optimal binarization threshold for BAA, further aligning training dynamics with inference behavior. Extensive experiments across various architectures and datasets demonstrate the effectiveness of our approach. Beyond ED, BAA offers a generalizable strategy for bridging the gap between continuous optimization and discrete evaluation in structured prediction tasks.

Wenjin Qin, Hailin Wang, Hao Shu et al.

In recent years, tensor decomposition-based approaches for hyperspectral anomaly detection (HAD) have gained significant attention in the field of remote sensing. However, existing methods often fail to fully leverage both the global correlations and local smoothness of the background components in hyperspectral images (HSIs), which exist in both the spectral and spatial domains. This limitation results in suboptimal detection performance. To mitigate this critical issue, we put forward a novel HAD method named HAD-EUNTRFR, which incorporates an enhanced unified nonconvex tensor ring (TR) factors regularization. In the HAD-EUNTRFR framework, the raw HSIs are first decomposed into background and anomaly components. The TR decomposition is then employed to capture the spatial-spectral correlations within the background component. Additionally, we introduce a unified and efficient nonconvex regularizer, induced by tensor singular value decomposition (TSVD), to simultaneously encode the low-rankness and sparsity of the 3-D gradient TR factors into a unique concise form. The above characterization scheme enables the interpretable gradient TR factors to inherit the low-rankness and smoothness of the original background. To further enhance anomaly detection, we design a generalized nonconvex regularization term to exploit the group sparsity of the anomaly component. To solve the resulting doubly nonconvex model, we develop a highly efficient optimization algorithm based on the alternating direction method of multipliers (ADMM) framework. Experimental results on several benchmark datasets demonstrate that our proposed method outperforms existing state-of-the-art (SOTA) approaches in terms of detection accuracy.

Xinkai Du, Quanjie Han, Chao Lv et al.

Open-domain Question Answering (QA) has garnered substantial interest by combining the advantages of faithfully retrieved passages and relevant passages generated through Large Language Models (LLMs). However, there is a lack of definitive labels available to pair these sources of knowledge. In order to address this issue, we propose an unsupervised and simple framework called Bi-Reranking for Merging Generated and Retrieved Knowledge (BRMGR), which utilizes re-ranking methods for both retrieved passages and LLM-generated passages. We pair the two types of passages using two separate re-ranking methods and then combine them through greedy matching. We demonstrate that BRMGR is equivalent to employing a bipartite matching loss when assigning each retrieved passage with a corresponding LLM-generated passage. The application of our model yielded experimental results from three datasets, improving their performance by +1.7 and +1.6 on NQ and WebQ datasets, respectively, and obtaining comparable result on TriviaQA dataset when compared to competitive baselines.