Signal Processing最新文献

{"title":"Exploring synergies: Advancing neuroscience with machine learning","authors":"Marzieh Ajirak ,&nbsp;Tülay Adali ,&nbsp;Saeid Sanei ,&nbsp;Logan Grosenick ,&nbsp;Petar M. Djurić","doi":"10.1016/j.sigpro.2025.110116","DOIUrl":"10.1016/j.sigpro.2025.110116","url":null,"abstract":"<div><div>Machine learning (ML) has transformed neuroscience research by providing powerful tools to analyze neural data, uncover brain connectivity, and guide therapeutic interventions. This paper presents core mathematical frameworks in ML that address critical challenges in neuroscience. We introduce state-space models for closed-loop neurostimulation and discrete representation learning methods that improve the interpretability of time-series analysis by extracting meaningful patterns from complex neural recordings. We also describe approaches for revealing inter-regional brain connectivity through high-dimensional time series analysis using Gaussian processes. In the context of multi-subject neuroimaging, we explore independent vector analysis to identify shared patterns that preserve individual differences. Finally, we examine distributed beamforming techniques to localize seizure sources from EEG data, an essential component of surgical planning for epilepsy treatment. These methodological innovations illustrate the growing role of ML in neuroscience via interpretable, adaptive, and personalized tools that analyze brain activity and support data-driven interventions.</div></div>","PeriodicalId":49523,"journal":{"name":"Signal Processing","volume":"238 ","pages":"Article 110116"},"PeriodicalIF":3.4,"publicationDate":"2025-06-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144243073","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Distributed robust information filter for Markov jump systems with outliers","authors":"Hongyu Zhu,&nbsp;Zhongliang Jing,&nbsp;Minzhe Li","doi":"10.1016/j.sigpro.2025.110107","DOIUrl":"10.1016/j.sigpro.2025.110107","url":null,"abstract":"<div><div>This paper studies the problem of distributed state estimation for linear Markov jump systems subject to state and measurement outliers. Initially, a robust local multiple model information filter is developed by integrating the traditional interactive multiple model framework with statistical similarity measure. This local filter is then extended to a distributed version via a diffusion-based information fusion strategy. Furthermore, based on stochastic stability theory, sufficient conditions are derived to ensure the boundedness of estimation errors in the mean square sense. The simulation results demonstrate the effectiveness of the proposed local and distributed multiple model filters in the presence of state and measurement outliers.</div></div>","PeriodicalId":49523,"journal":{"name":"Signal Processing","volume":"238 ","pages":"Article 110107"},"PeriodicalIF":3.4,"publicationDate":"2025-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144243072","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Sequential maximum-likelihood estimation of wideband polynomial-phase signals on sensor array","authors":"Kaleb Debre ,&nbsp;Tai Fei ,&nbsp;Marius Pesavento","doi":"10.1016/j.sigpro.2025.110105","DOIUrl":"10.1016/j.sigpro.2025.110105","url":null,"abstract":"<div><div>This paper presents a novel sequential estimator for the direction-of-arrival and polynomial coefficients of wideband polynomial-phase signals impinging on a sensor array. Addressing the computational challenges of maximum-likelihood estimation for this problem, we propose a method leveraging random sampling consensus (RANSAC) applied to the time-frequency spatial signatures of sources. Our approach supports multiple sources and higher-order polynomials by employing coherent array processing and sequential approximations of the maximum-likelihood cost function. We also propose a low-complexity variant that estimates source directions via angular domain random sampling. Numerical evaluations demonstrate that the proposed methods achieve Cramér-Rao bounds in challenging multi-source scenarios, including closely spaced time-frequency spatial signatures, highlighting their suitability for advanced radar signal processing applications.</div></div>","PeriodicalId":49523,"journal":{"name":"Signal Processing","volume":"238 ","pages":"Article 110105"},"PeriodicalIF":3.4,"publicationDate":"2025-05-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144243074","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Universum driven adaptive robust Adaboost twin extreme learning machine imbalance learning framework for pattern classification","authors":"Tian Tang ,&nbsp;Jun Ma ,&nbsp;Rongyu Qiao ,&nbsp;Xiaomei Sun","doi":"10.1016/j.sigpro.2025.110104","DOIUrl":"10.1016/j.sigpro.2025.110104","url":null,"abstract":"<div><div>Pattern recognition and machine learning research has demonstrated that Universum, as a third class distinct from the positive and negative classes, can be integrated with prior knowledge to improve the generalization performance of a model. This approach enables the incorporation of prior knowledge into the learning process, facilitating the development of more accurate models. This paper proposes a novel learning framework, termed the Universum-driven adaptive robust AdaBoost twin extreme learning machine imbalance learning framework (ARATELM), for addressing class imbalance classification problems. In this framework, a new generalized smooth uncapped adaptive robust loss function called <span><math><mrow><msub><mrow><mi>L</mi></mrow><mrow><mi>θ</mi></mrow></msub><mrow><mo>(</mo><mi>u</mi><mo>)</mo></mrow></mrow></math></span> is designed to improve the robustness of ARATELM. The generalized smooth uncapped adaptive robust loss function <span><math><mrow><msub><mrow><mi>L</mi></mrow><mrow><mi>θ</mi></mrow></msub><mrow><mo>(</mo><mi>u</mi><mo>)</mo></mrow></mrow></math></span> aims to address the problems caused by the capped loss function <span><math><msub><mrow><mi>L</mi></mrow><mrow><mi>θ</mi><mi>ɛ</mi></mrow></msub></math></span>, which ignores normal data points and introduces non-differentiability. Concurrently, <span><math><mrow><msub><mrow><mi>L</mi></mrow><mrow><mi>δ</mi></mrow></msub><mrow><mo>(</mo><mi>u</mi><mo>)</mo></mrow></mrow></math></span> effectively inherits the adaptability of <span><math><msub><mrow><mi>L</mi></mrow><mrow><mi>θ</mi><mi>ɛ</mi></mrow></msub></math></span> through the utilization of the adaptive parameter, <span><math><mi>δ</mi></math></span>, during the learning process. This enables the selection of diverse robust loss functions for different learning tasks, thereby enhancing the generalization performance of our method. Furthermore, the Universum data are taken into account in the proposed method, and prior information regarding the distribution of said data is provided; this enhances the generalization performance of the model. Additionally, the learning impact of our approach has been optimized through the integration of AdaBoost into ARATELM. Comprehensive experimental results across various class imbalance scenarios demonstrate that our presented method outperforms other methods in terms of robustness, classification accuracy, and other critical performance metrics.</div></div>","PeriodicalId":49523,"journal":{"name":"Signal Processing","volume":"238 ","pages":"Article 110104"},"PeriodicalIF":3.4,"publicationDate":"2025-05-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144203927","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Reinforced graph aggregation cross-domain few-shot learning for hyperspectral remote sensing image classification","authors":"Jingpeng Gao,&nbsp;Geng Chen,&nbsp;Xiangyu Ji,&nbsp;Chen Shen","doi":"10.1016/j.sigpro.2025.110101","DOIUrl":"10.1016/j.sigpro.2025.110101","url":null,"abstract":"<div><div>Few-shot learning (FSL) has been introduced to hyperspectral Image (HSI) classification due to the scarcity of labeled samples. Graph Neural Network (GNN) based FSL methods show excellent performance. Nevertheless, existing methods neglect that the graph data of different topological structures may require various aggregation iterations. It is difficult to extract global topological information in different domains with a fixed number of layers. A reinforced graph aggregation cross-domain FSL (RGA-CFSL) method is proposed, integrating FSL and deep reinforcement learning (DRL) into a unified framework. Specifically, supervised contrastive learning with multi-metric constraints (MSCL) is designed to provide stable prototypes. Meanwhile, we introduce a DRL model into a designed global information extraction (GIE) module, which alleviates domain shift at the topological structure level. The DRL model facilitates the extraction of global topological information, which dynamically predicts the optimal architectures of GNNs required for given graph data. Furthermore, an inter-scale feature fusion (IFF) module is designed to capture representative distribution information in domains and reduce domain shift at the distribution level, which aggregates global topological information and local spatial–spectral information. Experimental results on four target HSI datasets demonstrate the our RGA-CFSL obtains superior performance.</div></div>","PeriodicalId":49523,"journal":{"name":"Signal Processing","volume":"238 ","pages":"Article 110101"},"PeriodicalIF":3.4,"publicationDate":"2025-05-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144212545","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Image-to-Image Steganography based on multimodal generative model","authors":"Jingyuan Jiang,&nbsp;Zichi Wang,&nbsp;Xinpeng Zhang","doi":"10.1016/j.sigpro.2025.110106","DOIUrl":"10.1016/j.sigpro.2025.110106","url":null,"abstract":"<div><div>The diffusion-based multimodal generative model is highly popular and successful across various fields. It possesses significant potential within the domain of image steganography. Several researchers have proposed a model for image steganography that is based on Latent Diffusion Models (LDMs). Nevertheless, the proposed model exhibits low visual quality in the generated stego images, demonstrates limited capability in content control of the stego images via prompts, and low diversity of the generated stego images. To address the aforementioned issues, we propose a novel Image-to-Image Steganography Model(I2IStega) based on LDMs. The proposed I2IStega model excels in generating high-quality Stego images, efficiently utilizing the performance capabilities of LDMs. Furthermore, the model has markedly enhanced both the controllability and diversity of the content of the generated stego image. Compared with the previously proposed LDM-based image steganography models, I2IStega has a simpler model structure and less computational effort. The experimental results indicate that our method exhibits superior performance in terms of controllability, stego image quality, extract image quality, and security.</div></div>","PeriodicalId":49523,"journal":{"name":"Signal Processing","volume":"238 ","pages":"Article 110106"},"PeriodicalIF":3.4,"publicationDate":"2025-05-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144212544","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Low-complexity hybrid beamforming for multi-cell mmWave massive MIMO: A primitive Kronecker decomposition approach","authors":"Teng Sun ,&nbsp;Guangxu Zhu ,&nbsp;Xiaofan Li ,&nbsp;Jiancun Fan ,&nbsp;Minghua Xia","doi":"10.1016/j.sigpro.2025.110102","DOIUrl":"10.1016/j.sigpro.2025.110102","url":null,"abstract":"<div><div>To circumvent the high path loss of mmWave propagation and reduce the hardware cost of massive multiple-input multiple-output antenna systems, full-dimensional hybrid beamforming is critical in 5G and beyond wireless communications. Concerning an uplink multi-cell system with a large-scale uniform planar antenna array, this paper designs an efficient hybrid beamformer using primitive Kronecker decomposition and dynamic factor allocation, where the analog beamformer applies to null the inter-cell interference and simultaneously enhances the desired signals. In contrast, the digital beamformer mitigates the intra-cell interference using the minimum mean square error (MMSE) criterion. Then, due to the low accuracy of phase shifters inherent in the analog beamformer, a low-complexity hybrid beamformer is developed to slow its adjustment speed. Next, an optimality analysis from a subspace perspective is performed, and a sufficient condition for optimal antenna configuration is established. Finally, simulation results demonstrate that the achievable sum rate of the proposed beamformer approaches that of the optimal pure digital MMSE scheme, yet with much lower computational complexity and hardware cost.</div></div>","PeriodicalId":49523,"journal":{"name":"Signal Processing","volume":"238 ","pages":"Article 110102"},"PeriodicalIF":3.4,"publicationDate":"2025-05-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144196416","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A manifold-based codebook design scheme for near-field LoS channel in XL-MIMO systems","authors":"Zhen Yang,&nbsp;Tianbao Gao,&nbsp;Yunchao Song,&nbsp;Chen Liu","doi":"10.1016/j.sigpro.2025.110103","DOIUrl":"10.1016/j.sigpro.2025.110103","url":null,"abstract":"<div><div>In this paper, we propose a manifold-based codebook design scheme for near-field line-of-sight (LoS) channel in extremely large-scale multiple-input–multiple output (XL-MIMO) systems, where the user is equipped with multiple antennas. The LoS XL-MIMO channel model has been developed, considering the spatial distribution characteristics of multiple receiving antennas. The proposed scheme employs manifold optimization to design the codebook, resulting in a higher achievable rate compared to the conventional polar-domain codebook. Specifically, to establish the codebook, the angle and distance parameters in space are quantified by minimizing the incoherence between two channel matrices of different user locations. Since the near-field LoS channel matrix is no longer rank one, multiple codewords are designed for each quantified point to cover the LoS paths. Furthermore, the joint design of codewords at the BS and user side is formulated as an optimization problem with a constant modulus constraint, which defines a geometric structure in the form of a complex circle manifold. To solve the constant modulus optimization problem, we decouple it into two independent sub-problems. The Riemannian gradient descent algorithm on the complex circle manifold is applied to efficiently solve such problems. Numerical results demonstrate the superiority of our proposed scheme in improving the achievable rate.</div></div>","PeriodicalId":49523,"journal":{"name":"Signal Processing","volume":"238 ","pages":"Article 110103"},"PeriodicalIF":3.4,"publicationDate":"2025-05-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144190048","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Target localization with coprime multistatic MIMO radar via coupled canonical polyadic decomposition based on joint eigenvalue decomposition","authors":"Guo-Zhao Liao ,&nbsp;Xiao-Feng Gong ,&nbsp;Wei Liu ,&nbsp;Hing Cheung So","doi":"10.1016/j.sigpro.2025.110099","DOIUrl":"10.1016/j.sigpro.2025.110099","url":null,"abstract":"<div><div>This paper investigates target localization using a multistatic multiple-input multiple-output (MIMO) radar system with two distinct coprime array configurations: coprime L-shaped arrays and coprime planar arrays. The observed signals are modeled as tensors that admit a coupled canonical polyadic decomposition (C-CPD) model. For each configuration, a C-CPD method is presented based on joint eigenvalue decomposition (J-EVD). This computational framework includes (semi-)algebraic and optimization-based C-CPD algorithms and target localization that fuses direction-of-arrivals (DOAs) information to calculate the optimal position of each target. Specifically, the proposed (semi-)algebraic methods exploit the rotational invariance of the Vandermonde structure in coprime arrays, similar to the multiple invariance property of estimation of signal parameters via rotational invariance techniques (ESPRIT), which transforms the model into a J-EVD problem and reduces computational complexity. The study also investigates the working conditions of the algorithm to understand model identifiability. Additionally, the proposed method does not rely on prior knowledge of non-orthogonal probing waveforms and is effective in challenging underdetermined scenarios. Experimental results demonstrate that our method outperforms existing tensor-based approaches in both accuracy and computational efficiency.</div></div>","PeriodicalId":49523,"journal":{"name":"Signal Processing","volume":"238 ","pages":"Article 110099"},"PeriodicalIF":3.4,"publicationDate":"2025-05-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144203925","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Texture and geometric feature-fusion-based network for Dunhuang mural inpainting","authors":"Yutong Hou ,&nbsp;Shiqiang Du ,&nbsp;Huaikun Zhang ,&nbsp;Jizhao Liu ,&nbsp;Jinying Liu ,&nbsp;Jing Lian","doi":"10.1016/j.sigpro.2025.110096","DOIUrl":"10.1016/j.sigpro.2025.110096","url":null,"abstract":"<div><div>In recent years, deep-learning-based image inpainting methods have emerged as a popular research area. However, the application of such methods to mural inpainting presents several challenges. First, mural images often contain complex textures and rich details. Traditional inpainting methods struggle to preserve texture and detail information effectively and cannot ensure consistency between restored areas and the original mural. Second, the missing regions of murals often contain complex geometric structures and artistic styles, requiring mural inpainting algorithms to understand an image’s global semantics and accurately capture local details. This paper proposes a texture and geometric feature fusion network for Dunhuang mural inpainting consisting of two subnetworks: a Primary Inpainting Network (PIN) and a Refinement Enhancement Network (REN). The PIN extracts geometric features by incorporating fresco line drawings. It utilizes a Mamba-enhanced encoding module and gated convolution in its encoder to capture image texture features effectively, thereby enhancing the clarity of texture details. Then, the Dynamic Multi-scale Semantic Fusion Module (DMSFM) combines global and local information from texture and geometric features, completing the initial inpainting of a damaged mural. The REN specializes in inpainting image details by recovering complex textures, fine edges, and local structures. Randomly selected Narrative, Buddhist, and Caisson murals from a Dunhuang mural painting dataset were used to test the proposed network. Comparative experimental results demonstrate that the proposed method achieves superior mural inpainting outcomes compared with popular existing methods.</div></div>","PeriodicalId":49523,"journal":{"name":"Signal Processing","volume":"238 ","pages":"Article 110096"},"PeriodicalIF":3.4,"publicationDate":"2025-05-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144190047","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}