Signal Processing最新文献

{"title":"Understanding encoder–decoder structures in machine learning using information measures","authors":"Jorge F. Silva ,&nbsp;Victor Faraggi ,&nbsp;Camilo Ramirez ,&nbsp;Alvaro Egaña ,&nbsp;Eduardo Pavez","doi":"10.1016/j.sigpro.2025.109983","DOIUrl":"10.1016/j.sigpro.2025.109983","url":null,"abstract":"<div><div>We present a theory of representation learning to model and understand the role of encoder–decoder design in machine learning (ML) from an information-theoretic angle. We use two main information concepts, information sufficiency (IS) and mutual information loss to represent predictive structures in machine learning. Our first main result provides a functional expression that characterizes the class of probabilistic models consistent with an IS encoder–decoder latent predictive structure. This result formally justifies the encoder–decoder forward stages many modern ML architectures adopt to learn latent (compressed) representations for classification. To illustrate IS as a realistic and relevant model assumption, we revisit some known ML concepts and present some interesting new examples: invariant, robust, sparse, and digital models. Furthermore, our IS characterization allows us to tackle the fundamental question of how much performance could be lost, using the cross entropy risk, when a given encoder–decoder architecture is adopted in a learning setting. Here, our second main result shows that a mutual information loss quantifies the lack of expressiveness attributed to the choice of a (biased) encoder–decoder ML design. Finally, we address the problem of universal cross-entropy learning with an encoder–decoder design where necessary and sufficiency conditions are established to meet this requirement. In all these results, Shannon’s information measures offer new interpretations and explanations for representation learning.</div></div>","PeriodicalId":49523,"journal":{"name":"Signal Processing","volume":"234 ","pages":"Article 109983"},"PeriodicalIF":3.4,"publicationDate":"2025-03-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143621480","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 language-guided cross-modal semantic fusion retrieval method","authors":"Ligu Zhu ,&nbsp;Fei Zhou ,&nbsp;Suping Wang ,&nbsp;Lei Shi ,&nbsp;Feifei Kou ,&nbsp;Zeyu Li ,&nbsp;Pengpeng Zhou","doi":"10.1016/j.sigpro.2025.109993","DOIUrl":"10.1016/j.sigpro.2025.109993","url":null,"abstract":"<div><div>The rapid growth of the Internet and big data has led to the generation of large-scale multimodal data, presenting challenges for traditional retrieval methods. These methods often rely on a two-stage architecture involving retrieval and reranking, which struggles with integrating the semantic differences between visual and textual modalities. This limitation hampers the fusion of information and reduces the accuracy and efficiency of cross-modal retrieval. To overcome these challenges, we propose FusionRM, a language-guided cross-modal semantic fusion retrieval method. FusionRM utilizes the expressive power of textual semantics to bridge the knowledge gap between visual and linguistic modalities. By combining implicit visual knowledge with explicit textual knowledge, FusionRM creates a unified embedding space, aligning semantics across modalities and improving retrieval accuracy and efficiency of multimodal information processing. Experiments on the multi-hop, multimodal WebQA dataset show that FusionRM outperforms traditional methods across multiple metrics, demonstrating superior performance and generalization in open-domain retrieval.</div></div>","PeriodicalId":49523,"journal":{"name":"Signal Processing","volume":"234 ","pages":"Article 109993"},"PeriodicalIF":3.4,"publicationDate":"2025-03-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143621479","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":"Enhanced channel estimation for double RIS-aided MIMO systems using coupled tensor decompositions","authors":"Gerald C. Nwalozie ,&nbsp;André L.F. de Almeida ,&nbsp;Martin Haardt","doi":"10.1016/j.sigpro.2025.109979","DOIUrl":"10.1016/j.sigpro.2025.109979","url":null,"abstract":"<div><div>In this paper, we consider a double-RIS (D-RIS)-aided flat-fading MIMO system and propose an interference-free channel training and estimation protocol, where the two single-reflection links and the one double-reflection link are estimated separately. Specifically, by using the proposed training protocol, the signal measurements of a particular reflection link can be extracted interference-free from the measurements of the superposition of the three links. We show that some channels are associated with two different components of the received signal.Exploiting the common channels involved in the single and double reflection links while recasting the received signals as tensors, we formulate the coupled tensor-based least square Khatri–Rao factorization (C-KRAFT) algorithm which is a closed-form solution and an enhanced iterative solution with less restrictions on the identifiability constraints, the coupled-alternating least square (C-ALS) algorithm. The C-KRAFT and C-ALS based channel estimation schemes are used to obtain the channel matrices in both single and double reflection links.We show that the proposed coupled tensor decomposition-based channel estimation schemes offer more accurate channel estimates under less restrictive identifiability constraints compared to competing channel estimation methods. Simulation results are provided showing the effectiveness of the proposedalgorithms.</div></div>","PeriodicalId":49523,"journal":{"name":"Signal Processing","volume":"234 ","pages":"Article 109979"},"PeriodicalIF":3.4,"publicationDate":"2025-03-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143621481","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A novel adaptive direct signal interference cancellation method for underwater active electromagnetic detection systems using an error signal power-based variable step-size Fast Block Least Mean Square algorithm","authors":"Yan Ma,&nbsp;Hao Zhang,&nbsp;Ke Yang","doi":"10.1016/j.sigpro.2025.109957","DOIUrl":"10.1016/j.sigpro.2025.109957","url":null,"abstract":"<div><div>As a promising candidate to detect the underwater targets, the underwater active electromagnetic detection system has gained more attention in the field of marine exploration. Because of the separation of the transmitting and receiving antennas, the direct signal interference in the receiving end would be a great challenge during the operation, which severely deteriorates the system’s capability of target identification and localization. This paper proposes a novel direct signal interference cancellation method based on an error signal power-driven variable step-size fast block Least Mean Square (PVSS-FBLMS) algorithm to address the direct signal interference cancellation issue in the underwater active EM detection system. The proposed algorithm aligns well with the system’s characteristic of block-wise data transmission and demonstrates lower complexity, especially for higher FIR filter orders, compared to the Least Mean Square and Normalized Least Mean Square algorithms. Considering the limited time available for direct signal interference cancellation, the proposed algorithm introduces a variable step-size adjustment criterion based on the error signal power to accelerate convergence, while ensuring the robustness of the algorithm. The simulation has been conducted, which demonstrates that the proposed algorithm converges approximate 46.8% faster than that of the improved variable step-size least mean square (IVSS-LMS) and variable step-size least mean square (VSS-LMS) algorithms while its computational complexity is only about 40% that of the two algorithms when the FIR filter order is 128. The direct signal interference cancellation performance of the proposed algorithm is significantly better than that of the IVSS-LMS and VSS-LMS algorithms at low DSI-to-noise ratios under the simulation condition. Additionally, the simulation results show that the proposed algorithm performs steadily in a highly abrupt change of the noise and the DSI’s amplitude. Besides, the in-field experiments are conducted to validate the effectiveness of the proposed algorithm. The experimental results show that the convergence rate of the proposed algorithm is notably faster than that of the IVSS-LMS and VSS-LMS algorithms, with a steady-state output variance on the order of <span><math><mrow><mn>1</mn><msup><mrow><mn>0</mn></mrow><mrow><mo>−</mo><mn>6</mn></mrow></msup><msup><mrow><mi>V</mi></mrow><mrow><mn>2</mn></mrow></msup></mrow></math></span>.</div></div>","PeriodicalId":49523,"journal":{"name":"Signal Processing","volume":"234 ","pages":"Article 109957"},"PeriodicalIF":3.4,"publicationDate":"2025-03-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143591982","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":"Characterization of irregular cyclicities in heavy-tailed data","authors":"Antonio Napolitano ,&nbsp;Agnieszka Wyłomańska","doi":"10.1016/j.sigpro.2025.109980","DOIUrl":"10.1016/j.sigpro.2025.109980","url":null,"abstract":"<div><div>The statistical characterization of heavy-tailed data with hidden irregular periodicities is introduced. Specifically, processes generated by the interaction of random phenomena with heavy-tailed distribution and almost-periodic phenomena with possibly irregular or disturbed peridicities are characterized in terms of fractional lower-order moments. For this wide class of processes, fractional lower-order moments are expressed as the superposition of amplitude- and angle-modulated sine waves. The model introduced in the paper extends the previously introduced one for heavy-tailed almost cyclostationary (ACS) processes to the case of irregular periodicities. Moreover, it introduces for the class of the oscillatory ACS processes a characterization in terms of fractional lower-order moments. For the new class, the problem of statistical function estimation is addressed. The effectiveness of the proposed methodology is corroborated by the analysis of simulated alpha-stable time-warped ACS processes and of real acoustic helicopter data.</div></div>","PeriodicalId":49523,"journal":{"name":"Signal Processing","volume":"234 ","pages":"Article 109980"},"PeriodicalIF":3.4,"publicationDate":"2025-03-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143591999","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Elliptical Wishart distributions: Information geometry, maximum likelihood estimator, performance analysis and statistical learning","authors":"Imen Ayadi,&nbsp;Florent Bouchard,&nbsp;Frédéric Pascal","doi":"10.1016/j.sigpro.2025.109981","DOIUrl":"10.1016/j.sigpro.2025.109981","url":null,"abstract":"<div><div>This paper deals with Elliptical Wishart distributions – which generalize the Wishart distribution – in the context of signal processing and machine learning. Two algorithms to compute the maximum likelihood estimator (MLE) are proposed: a fixed point algorithm and a Riemannian optimization method based on the derived information geometry of Elliptical Wishart distributions. The existence and uniqueness of the MLE are characterized as well as the convergence of both estimation algorithms. Statistical properties of the MLE are also investigated such as consistency, asymptotic normality and an intrinsic version of Fisher efficiency. On the statistical learning side, novel classification and clustering methods are designed. For the <span><math><mi>t</mi></math></span>-Wishart distribution, the performance of the MLE and statistical learning algorithms are evaluated on both simulated and real EEG and hyperspectral data, showcasing the interest of our proposed methods.</div></div>","PeriodicalId":49523,"journal":{"name":"Signal Processing","volume":"234 ","pages":"Article 109981"},"PeriodicalIF":3.4,"publicationDate":"2025-03-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143654604","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":"Phase differences estimation for diagonal ULAs within a fully filled rectangular array based on degenerated spatial ARMA process","authors":"Guijin Yao,&nbsp;Jiye Liu,&nbsp;Hairong Zhang,&nbsp;Yue Li","doi":"10.1016/j.sigpro.2025.109985","DOIUrl":"10.1016/j.sigpro.2025.109985","url":null,"abstract":"<div><div>A generalized estimation method of diagonal phase differences of external sources incident upon a fully-filled rectangular array (FFRA) is proposed based on degenerate spatial <em>ARMA</em> process. Various diagonal uniform linear arrays (ULAs) within FFRAs are first classified by function forms of diagonal phase differences, the ULAs sharing the same diagonal phase difference belong to one category. The modified Yule–Walker (MYW) system of linear equations and the root-finding polynomial are first derived for FFRA ULAs. Owing to diagonal interspacings larger than half of carrier wavelength, ambiguity problem of diagonal phase differences has arisen in estimation. Utilizing the explicit linear-combination relationships satisfied by diagonal and axial phase differences, a simple and effective elimination scheme of estimate ambiguity of diagonal phase differences is proposed in which actual intervals of no ambiguity are deduced by making use of the estimates of axial phase differences. With different FFRA diagonal ULAs on <span><math><mrow><mi>X</mi><mo>−</mo><mi>Y</mi></mrow></math></span> sensor plane, it is numerically manifested by Monte-Carlo trials that the proposed method is effective for both independent and coherent sources and the Root mean square errors (RMSEs) are slowly convergent to the corresponding Cramer–Rao bounds (CRBs) after estimate ambiguities are eliminated. The consistency of estimation performance for diagonal ULAs belonging to one category is exhibited by their identical RMSEs. Because of the ability to exploit axial and diagonal ULAs, the proposed estimation method provides the basis of two-dimensional direction of arrival (2-D DoA) estimation with ULA combinations of FFRAs.</div></div>","PeriodicalId":49523,"journal":{"name":"Signal Processing","volume":"234 ","pages":"Article 109985"},"PeriodicalIF":3.4,"publicationDate":"2025-03-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143601400","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":"EscapeTrack: Multi-object tracking with estimated camera parameters","authors":"Kefu Yi ,&nbsp;Hao Wu ,&nbsp;Wei Hao ,&nbsp;Rongdong Hu","doi":"10.1016/j.sigpro.2025.109958","DOIUrl":"10.1016/j.sigpro.2025.109958","url":null,"abstract":"<div><div>Multi-object tracking (MOT) remains a challenging task in dynamic environments. While most 2D tracking methods focus solely on the image plane, they often neglect the Ground Plane Assumption (GPA) — the principle that targets typically move on a consistent ground plane. This is because camera parameters are difficult to obtain and are not very reliable in scenarios involving camera motion or where the GPA does not apply. To address this issue, we propose EscapeTrack, a novel MOT algorithm that robustly handles imprecise camera parameters. Unlike conventional homography projection methods prone to calibration errors, EscapeTrack innovatively models target coordinates on the ground plane as latent variables within a Kalman filter framework. By constructing an observation model that projects these latent states onto the image plane, our method achieves superior tracking accuracy even with significant parameter noise. Extensive evaluations demonstrate state-of-the-art performance on MOT17, MOT20, DanceTrack, SportsMOT, and BDD100K benchmarks. Notably, EscapeTrack excels in scenarios with camera motion or GPA violations, by inherently treating such cases as camera parameter estimation errors. This robustness enables practical deployment in real-world systems where precise calibration is infeasible, advancing intelligent tracking in complex dynamic environments. The source code will be available at <span><span>https://github.com/corfyi/EscapeTrack</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":49523,"journal":{"name":"Signal Processing","volume":"234 ","pages":"Article 109958"},"PeriodicalIF":3.4,"publicationDate":"2025-03-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143601401","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 reduced-order Kalman consensus filter for multisensor networked descriptor systems","authors":"Minghu Zhang,&nbsp;Shuli Sun","doi":"10.1016/j.sigpro.2025.109991","DOIUrl":"10.1016/j.sigpro.2025.109991","url":null,"abstract":"<div><div>In the context of multisensor linear discrete networked descriptor systems, an equivalence transformation, achieved via singular value decomposition, leads to the derivation of two lower-dimensional non-descriptor subsystems. Each network node can perform state estimation based on data of its own and its neighboring nodes. Applying the Kalman consensus filter (KCF) framework, wherein one-step prediction estimates of a reduced-order subsystem are exchanged among network nodes, a distributed reduced-order KCF is designed for each sensor node, incorporating multiple consensus gains. This design facilitates collaborative state estimation by enabling nodes to leverage both their own measurements and the prediction estimates received from their neighbors. The optimal Kalman filtering gains and the optimal consensus filtering gains are determined by minimizing the trace of the filtering error covariance matrix. The investigation delves into the stability and steady-state characteristics of the tailored distributed reducer-order filtering systems. The performance of the algorithms is confirmed through illustrative simulation cases.</div></div>","PeriodicalId":49523,"journal":{"name":"Signal Processing","volume":"234 ","pages":"Article 109991"},"PeriodicalIF":3.4,"publicationDate":"2025-03-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143591983","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":"Data-reuse recursive least-squares algorithm with Riemannian manifold constraint","authors":"Haiquan Zhao,&nbsp;Haolin Wang,&nbsp;Yi Peng","doi":"10.1016/j.sigpro.2025.109982","DOIUrl":"10.1016/j.sigpro.2025.109982","url":null,"abstract":"<div><div>Actual signals often contain nonlinear manifold structures, but traditional filtering algorithms assume data are embedded in Euclidean space, which makes them less effective when handling complicated noise and manifold data. To address these challenges, Riemannian geometry constraints to the traditional data-reuse recursive least-squares (DR-RLS) algorithm is proposed in this paper. Therefore, a novel adaptive filtering algorithm combining the DR-RLS algorithm with Riemannian manifolds is proposed. This algorithm constrains the filter update process on the Riemannian manifold through exponential mapping, enabling better adaptation to nonlinear manifold data structures. Additionally, the tracking performance and convergence speed of the algorithm are enhanced by data reuse. The convergence and computational complexity of the proposed algorithm on the Riemannian manifold are also analyzed. Finally, the effectiveness of the proposed algorithm relative to other methods is demonstrated through simulation results.</div></div>","PeriodicalId":49523,"journal":{"name":"Signal Processing","volume":"234 ","pages":"Article 109982"},"PeriodicalIF":3.4,"publicationDate":"2025-03-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143641719","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}