Signal Processing最新文献_第9页

A novel noise-robust and lightweight underwater acoustic target recognition method based on BSCQT and DCAM 一种基于BSCQT和DCAM的新型抗噪轻量级水声目标识别方法

IF 3.6 2区工程技术 Q2 ENGINEERING, ELECTRICAL & ELECTRONIC

Signal Processing

Pub Date : 2026-04-01 Epub Date: 2025-11-06 DOI: 10.1016/j.sigpro.2025.110386

Qiyang Xiao , Yu Li , Xiaodong Zhai , Wenlong Jiang , Yong Jin , Ke Yuan , Wentao Shi

To address the challenges of low recognition accuracy and high computational overhead in noisy underwater environments, this paper proposes a novel noise-robust and lightweight underwater acoustic target recognition method based on Band-Specific Constant Q Transform (BSCQT) and Dynamic Context-Aware Masking (DCAM). First, BSCQT achieves effective noise suppression and feature extraction through multi-band adaptive weighting and feature concatenation. Then, by combining frequency-adaptive pooling granularity with traditional lightweight context-aware masking, a dynamic context-aware masking (DCAM) mechanism is constructed to implement adaptive attention on BSCQT features, improving recognition accuracy while maintaining low computational complexity. Furthermore, a Dynamic Context-Aware Masking Network (DCAMNet) is developed based on DCAM for hierarchical feature learning, integrating cascaded DCAM dense TDNN blocks for efficient information transmission. Finally, within the DCAMNet architecture, target recognition is accomplished through global pooling and fully connected classification layers. Extensive experimental results demonstrate that the proposed method achieves 99.23% recognition accuracy with only 0.55G Floating Point Operations (FLOPs) computational complexity, showing significant improvement in recognition efficiency compared to existing state-of-the-art methods and verifying the effectiveness of our approach.

针对水下噪声环境下识别精度低、计算量大的问题，提出了一种基于频带特定常数Q变换（BSCQT）和动态上下文感知掩蔽（DCAM）的轻型噪声鲁棒水声目标识别方法。首先，BSCQT通过多波段自适应加权和特征拼接实现了有效的噪声抑制和特征提取。然后，将频率自适应池化粒度与传统轻量级上下文感知掩蔽相结合，构建动态上下文感知掩蔽（DCAM）机制，实现对BSCQT特征的自适应关注，在保持较低计算复杂度的同时提高识别精度。在此基础上，开发了一种基于DCAM的动态上下文感知掩蔽网络（DCAMNet），用于分层特征学习，并集成了级联DCAM密集TDNN块，实现了高效的信息传输。最后，在DCAMNet体系结构中，通过全局池化和完全连接的分类层来完成目标识别。大量的实验结果表明，该方法在0.55G浮点运算（FLOPs）的计算复杂度下，识别准确率达到99.23%，与现有的最先进方法相比，识别效率有了显著提高，验证了该方法的有效性。

{"title":"A novel noise-robust and lightweight underwater acoustic target recognition method based on BSCQT and DCAM","authors":"Qiyang Xiao , Yu Li , Xiaodong Zhai , Wenlong Jiang , Yong Jin , Ke Yuan , Wentao Shi","doi":"10.1016/j.sigpro.2025.110386","DOIUrl":"10.1016/j.sigpro.2025.110386","url":null,"abstract":"<div><div>To address the challenges of low recognition accuracy and high computational overhead in noisy underwater environments, this paper proposes a novel noise-robust and lightweight underwater acoustic target recognition method based on Band-Specific Constant Q Transform (BSCQT) and Dynamic Context-Aware Masking (DCAM). First, BSCQT achieves effective noise suppression and feature extraction through multi-band adaptive weighting and feature concatenation. Then, by combining frequency-adaptive pooling granularity with traditional lightweight context-aware masking, a dynamic context-aware masking (DCAM) mechanism is constructed to implement adaptive attention on BSCQT features, improving recognition accuracy while maintaining low computational complexity. Furthermore, a Dynamic Context-Aware Masking Network (DCAMNet) is developed based on DCAM for hierarchical feature learning, integrating cascaded DCAM dense TDNN blocks for efficient information transmission. Finally, within the DCAMNet architecture, target recognition is accomplished through global pooling and fully connected classification layers. Extensive experimental results demonstrate that the proposed method achieves 99.23% recognition accuracy with only 0.55G Floating Point Operations (FLOPs) computational complexity, showing significant improvement in recognition efficiency compared to existing state-of-the-art methods and verifying the effectiveness of our approach.</div></div>","PeriodicalId":49523,"journal":{"name":"Signal Processing","volume":"241 ","pages":"Article 110386"},"PeriodicalIF":3.6,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145528910","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}

引用次数: 0

A space-time blind adaptive anti-interference method for distributed GNSS antenna arrays 分布式GNSS天线阵的空时盲自适应抗干扰方法

IF 3.6 2区工程技术 Q2 ENGINEERING, ELECTRICAL & ELECTRONIC

Signal Processing

Pub Date : 2026-04-01 Epub Date: 2025-11-13 DOI: 10.1016/j.sigpro.2025.110409

Binbin Ren , Song Li , Feiqiang Chen , Chenxuan Liu , Shaojie Ni

Antenna arrays and spatial adaptive processing are among the most effective means to enhance the anti-interference capability of Global Navigation Satellite System (GNSS) receivers. This paper proposes a novel space-time blind adaptive anti-interference method for a new spatially distributed GNSS antenna array. Compared to conventional spatial-only processing, space-time processing compensates for baseband delays caused by large inter-element spacing, achieving superior interference suppression. However, traditional blind algorithms often attenuate satellite signals due to the lack of gain constraints, reducing tracking and measurement precision. To address this, the principle of noncoherent channel combination is extended to space-time array processing to improve GNSS tracking precision. Since the original two-stage framework was designed for compact arrays and spatial-only processing, it does not address the code phase inconsistencies introduced by space-time processing in distributed arrays. A channel selection stage is thus introduced. The framework comprises three stages: interference suppression, optimal channel selection, and noncoherent channel combination. The method retains the advantages of blind processing and requires no external auxiliary information. Comparative analysis with state-of-the-art space-time blind anti-interference approaches demonstrates its effectiveness. Simulations confirm superior tracking performance and interference mitigation, validating its advantages over existing techniques.

天线阵列和空间自适应处理是提高全球导航卫星系统（GNSS）接收机抗干扰能力的最有效手段之一。针对一种新型空间分布式GNSS天线阵列，提出了一种新的空时盲自适应抗干扰方法。与传统的纯空间处理相比，时空处理补偿了由于元间间距大造成的基带延迟，实现了更好的干扰抑制。然而，传统的盲算法由于缺乏增益约束，往往使卫星信号衰减，降低了跟踪和测量精度。为解决这一问题，将非相干信道组合原理扩展到空时阵列处理中，提高GNSS跟踪精度。由于最初的两阶段框架是为紧凑数组和仅空间处理而设计的，因此它没有解决分布式数组中由时空处理引入的代码相位不一致性。这样就引入了信道选择阶段。该框架包括三个阶段：干扰抑制、最优信道选择和非相干信道组合。该方法保留了盲处理的优点，不需要外部辅助信息。通过与现有时空盲抗干扰方法的对比分析，验证了该方法的有效性。仿真验证了其优越的跟踪性能和抗干扰能力，验证了其优于现有技术的优势。

{"title":"A space-time blind adaptive anti-interference method for distributed GNSS antenna arrays","authors":"Binbin Ren , Song Li , Feiqiang Chen , Chenxuan Liu , Shaojie Ni","doi":"10.1016/j.sigpro.2025.110409","DOIUrl":"10.1016/j.sigpro.2025.110409","url":null,"abstract":"<div><div>Antenna arrays and spatial adaptive processing are among the most effective means to enhance the anti-interference capability of Global Navigation Satellite System (GNSS) receivers. This paper proposes a novel space-time blind adaptive anti-interference method for a new spatially distributed GNSS antenna array. Compared to conventional spatial-only processing, space-time processing compensates for baseband delays caused by large inter-element spacing, achieving superior interference suppression. However, traditional blind algorithms often attenuate satellite signals due to the lack of gain constraints, reducing tracking and measurement precision. To address this, the principle of noncoherent channel combination is extended to space-time array processing to improve GNSS tracking precision. Since the original two-stage framework was designed for compact arrays and spatial-only processing, it does not address the code phase inconsistencies introduced by space-time processing in distributed arrays. A channel selection stage is thus introduced. The framework comprises three stages: interference suppression, optimal channel selection, and noncoherent channel combination. The method retains the advantages of blind processing and requires no external auxiliary information. Comparative analysis with state-of-the-art space-time blind anti-interference approaches demonstrates its effectiveness. Simulations confirm superior tracking performance and interference mitigation, validating its advantages over existing techniques.</div></div>","PeriodicalId":49523,"journal":{"name":"Signal Processing","volume":"241 ","pages":"Article 110409"},"PeriodicalIF":3.6,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145579651","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}

引用次数: 0

Semi-supervised non-negative matrix factorization with weighted label propagation for data representation 基于加权标签传播的半监督非负矩阵分解

IF 3.6 2区工程技术 Q2 ENGINEERING, ELECTRICAL & ELECTRONIC

Signal Processing

Pub Date : 2026-04-01 Epub Date: 2025-11-11 DOI: 10.1016/j.sigpro.2025.110407

Wenjing Jing , Linzhang Lu , Weihua Ou

Label propagation has been widely used to enhance performance for clustering. Many semi-supervised non-negative matrix factorization (NMF) methods based on label propagation have been proposed. However, these methods mainly pay attention to learning a label prediction matrix, neglecting the efficient learning of a low-dimensional representation of original data. Additionally, they lead to inconsistent structures with NMF when leveraging label constraints, compromising the learning performance for low-dimensional representation and basis matrix. To address these problems, this paper proposes a novel semi-supervised NMF method named semi-supervised non-negative matrix factorization with weighted label propagation (SNMFWLP). Firstly, SNMFWLP considers an orthogonal constraint on basis matrix to minimize the redundancy in the process of decomposition in NMF. Secondly, SNMFWLP introduces a weighted label propagation model into NMF to learn an efficient low-dimensional representation used as label prediction matrix. The weighted label propagation model not only propagates label information but also maintains the structures consistent with structures of NMF, beneficial to a consistent low-dimensional representation. Additionally, effective algorithm and convergence analysis are also presented. Finally, numerous experiments on real-world data sets are conducted to demonstrate the superiority of the proposed method in comparison to several state-of-the-art unsupervised and semi-supervised NMF methods.

标签传播被广泛用于提高聚类的性能。人们提出了许多基于标签传播的半监督非负矩阵分解（NMF）方法。然而，这些方法主要关注标签预测矩阵的学习，忽略了对原始数据低维表示的有效学习。此外，当利用标签约束时，它们会导致与NMF结构不一致，从而影响低维表示和基矩阵的学习性能。针对这些问题，本文提出了一种新的半监督非负矩阵分解加权标签传播方法（SNMFWLP）。首先，SNMFWLP考虑基矩阵的正交约束，以最小化NMF分解过程中的冗余。其次，SNMFWLP在NMF中引入加权标签传播模型，学习高效的低维表示作为标签预测矩阵。加权标签传播模型不仅传播标签信息，而且保持了与NMF结构一致的结构，有利于一致的低维表示。并给出了有效的算法和收敛性分析。最后，在真实世界的数据集上进行了大量实验，以证明与几种最先进的无监督和半监督NMF方法相比，所提出的方法具有优越性。

{"title":"Semi-supervised non-negative matrix factorization with weighted label propagation for data representation","authors":"Wenjing Jing , Linzhang Lu , Weihua Ou","doi":"10.1016/j.sigpro.2025.110407","DOIUrl":"10.1016/j.sigpro.2025.110407","url":null,"abstract":"<div><div>Label propagation has been widely used to enhance performance for clustering. Many semi-supervised non-negative matrix factorization (NMF) methods based on label propagation have been proposed. However, these methods mainly pay attention to learning a label prediction matrix, neglecting the efficient learning of a low-dimensional representation of original data. Additionally, they lead to inconsistent structures with NMF when leveraging label constraints, compromising the learning performance for low-dimensional representation and basis matrix. To address these problems, this paper proposes a novel semi-supervised NMF method named semi-supervised non-negative matrix factorization with weighted label propagation (SNMFWLP). Firstly, SNMFWLP considers an orthogonal constraint on basis matrix to minimize the redundancy in the process of decomposition in NMF. Secondly, SNMFWLP introduces a weighted label propagation model into NMF to learn an efficient low-dimensional representation used as label prediction matrix. The weighted label propagation model not only propagates label information but also maintains the structures consistent with structures of NMF, beneficial to a consistent low-dimensional representation. Additionally, effective algorithm and convergence analysis are also presented. Finally, numerous experiments on real-world data sets are conducted to demonstrate the superiority of the proposed method in comparison to several state-of-the-art unsupervised and semi-supervised NMF methods.</div></div>","PeriodicalId":49523,"journal":{"name":"Signal Processing","volume":"241 ","pages":"Article 110407"},"PeriodicalIF":3.6,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145528906","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}

引用次数: 0

A novel sparse adaptive filter for suppressing impulsive disturbance in audio signals 一种抑制音频信号脉冲干扰的稀疏自适应滤波器

IF 3.6 2区工程技术 Q2 ENGINEERING, ELECTRICAL & ELECTRONIC

Signal Processing

Pub Date : 2026-04-01 Epub Date: 2025-11-07 DOI: 10.1016/j.sigpro.2025.110390

Lei Zhou , Hongqing Liu , Lu Gan , Yi Zhou , Maciej Niedźwiecki , Trieu-Kien Truong

This work studies the sparse adaptive filter designs for audio signal recovery under impulsive disturbance. By exploiting the sparse representation of desired signal and compressibility of impulsive disturbance, a joint sparse least mean

p

-norm (JSLMP) optimization, in which

ℓ_{p}

-norm (

1 \leq p \leq 2

) measures the data fidelity and

ℓ_{q}

-norm (

0 \leq q \leq 1

) enforces sparse solutions, is developed, termed as

ℓ_{q}

-JSLMP. The filter weights update is derived using gradient descent, and the Adam and variable step size (VSS) are integrated to accelerate convergence and avoid potential local minima. For the special case of

q = 1

, namely

ℓ_{1}

-JSLMP, its convergence condition and mean square deviation (MSD) analysis are derived. Finally, an application framework for processing corrupted audio signals is developed. Extensive experiments are conducted on both synthetic and real-measured impulsive noise data, comparing the proposed method with traditional algorithms as well as the deep learning-based GTCRN model. Results demonstrate that the proposed method yields superior perceptual quality and significantly lower memory consumption compared to GTCRN under impulsive disturbance.

本文研究了用于脉冲干扰下音频信号恢复的稀疏自适应滤波器设计。利用期望信号的稀疏表示和脉冲扰动的可压缩性，提出了一种联合稀疏最小平均p-范数（1≤p≤2）度量数据保真度，而q-范数（0≤q≤1）强制执行稀疏解的联合稀疏最小平均p-范数优化方法，称为q-JSLMP。采用梯度下降法推导滤波器权值更新，并将Adam和变步长（VSS）相结合，加快收敛速度，避免潜在的局部极小值。对于q=1的特殊情况，即1- jslmp，给出了其收敛条件和均方偏差（MSD）分析。最后，开发了一个用于处理音频损坏信号的应用框架。在合成和实测脉冲噪声数据上进行了大量的实验，将所提出的方法与传统算法以及基于深度学习的GTCRN模型进行了比较。结果表明，与脉冲干扰下的GTCRN相比，该方法具有更好的感知质量和更低的记忆消耗。

{"title":"A novel sparse adaptive filter for suppressing impulsive disturbance in audio signals","authors":"Lei Zhou , Hongqing Liu , Lu Gan , Yi Zhou , Maciej Niedźwiecki , Trieu-Kien Truong","doi":"10.1016/j.sigpro.2025.110390","DOIUrl":"10.1016/j.sigpro.2025.110390","url":null,"abstract":"<div><div>This work studies the sparse adaptive filter designs for audio signal recovery under impulsive disturbance. By exploiting the sparse representation of desired signal and compressibility of impulsive disturbance, a joint sparse least mean <span><math><mi>p</mi></math></span>-norm (JSLMP) optimization, in which <span><math><msub><mrow><mi>ℓ</mi></mrow><mrow><mi>p</mi></mrow></msub></math></span>-norm (<span><math><mrow><mn>1</mn><mo>≤</mo><mi>p</mi><mo>≤</mo><mn>2</mn></mrow></math></span>) measures the data fidelity and <span><math><msub><mrow><mi>ℓ</mi></mrow><mrow><mi>q</mi></mrow></msub></math></span>-norm (<span><math><mrow><mn>0</mn><mo>≤</mo><mi>q</mi><mo>≤</mo><mn>1</mn></mrow></math></span>) enforces sparse solutions, is developed, termed as <span><math><msub><mrow><mi>ℓ</mi></mrow><mrow><mi>q</mi></mrow></msub></math></span>-JSLMP. The filter weights update is derived using gradient descent, and the <em>Adam</em> and variable step size (VSS) are integrated to accelerate convergence and avoid potential local minima. For the special case of <span><math><mrow><mi>q</mi><mo>=</mo><mn>1</mn></mrow></math></span>, namely <span><math><msub><mrow><mi>ℓ</mi></mrow><mrow><mn>1</mn></mrow></msub></math></span>-JSLMP, its convergence condition and mean square deviation (MSD) analysis are derived. Finally, an application framework for processing corrupted audio signals is developed. Extensive experiments are conducted on both synthetic and real-measured impulsive noise data, comparing the proposed method with traditional algorithms as well as the deep learning-based GTCRN model. Results demonstrate that the proposed method yields superior perceptual quality and significantly lower memory consumption compared to GTCRN under impulsive disturbance.</div></div>","PeriodicalId":49523,"journal":{"name":"Signal Processing","volume":"241 ","pages":"Article 110390"},"PeriodicalIF":3.6,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145528912","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}

引用次数: 0

A structure adaptivity variation-based segmentation model for image with retinex and noise 基于结构自适应变化的视网膜和噪声图像分割模型

IF 3.6 2区工程技术 Q2 ENGINEERING, ELECTRICAL & ELECTRONIC

Signal Processing

Pub Date : 2026-04-01 Epub Date: 2025-11-13 DOI: 10.1016/j.sigpro.2025.110410

Yuan Wang , Tieyong Zeng , Zhi-Feng Pang , Hong Ge

Images often suffer from intensity inhomogeneity and noise during the imaging process, which poses considerable challenges for image segmentation. This paper proposes a novel adaptive structure variational model integrating Retinex theory for image segmentation, which effectively addresses intensity inhomogeneity and noise in the segmentation process. Grounded in Retinex theory, we decompose the image and impose two constraints: a piecewise-constant constraint on the reflectance to delineate homogeneous regions amid inhomogeneity, and a spatial smoothness constraint on the illumination to model the bias field. The primary novelty of the work lies in the introduction of an adaptive weighted matrix, comprising a rotation matrix and a scaling matrix, coupled with the gradient operator. This design enables the proposed model to perform anisotropic regularization, allowing it to adaptively capture directional structural features, preserve complex boundaries and textures, and simultaneously suppress noise effectively. We establish the existence of a solution for the proposed model and employ an efficient alternating minimization algorithm for numerical solution. Numerical experiments on synthetic, natural, and medical images demonstrate the desirable performance of the proposed model.

图像在成像过程中往往存在强度不均匀性和噪声，这对图像分割提出了相当大的挑战。结合Retinex理论，提出一种新的自适应结构变分模型用于图像分割，有效地解决了图像分割过程中的强度不均匀性和噪声问题。基于Retinex理论，我们对图像进行分解，并施加两个约束：对反射率进行分段常数约束，以在非均匀性中描绘均匀区域；对光照进行空间平滑约束，以模拟偏置场。这项工作的主要新颖之处在于引入了一个自适应加权矩阵，包括旋转矩阵和缩放矩阵，再加上梯度算子。该设计使所提出的模型能够进行各向异性正则化，使其能够自适应捕获定向结构特征，保留复杂的边界和纹理，同时有效地抑制噪声。我们建立了该模型解的存在性，并采用了一种高效的交替极小化算法求解数值解。在合成图像、自然图像和医学图像上的数值实验证明了该模型的良好性能。

{"title":"A structure adaptivity variation-based segmentation model for image with retinex and noise","authors":"Yuan Wang , Tieyong Zeng , Zhi-Feng Pang , Hong Ge","doi":"10.1016/j.sigpro.2025.110410","DOIUrl":"10.1016/j.sigpro.2025.110410","url":null,"abstract":"<div><div>Images often suffer from intensity inhomogeneity and noise during the imaging process, which poses considerable challenges for image segmentation. This paper proposes a novel adaptive structure variational model integrating Retinex theory for image segmentation, which effectively addresses intensity inhomogeneity and noise in the segmentation process. Grounded in Retinex theory, we decompose the image and impose two constraints: a piecewise-constant constraint on the reflectance to delineate homogeneous regions amid inhomogeneity, and a spatial smoothness constraint on the illumination to model the bias field. The primary novelty of the work lies in the introduction of an adaptive weighted matrix, comprising a rotation matrix and a scaling matrix, coupled with the gradient operator. This design enables the proposed model to perform anisotropic regularization, allowing it to adaptively capture directional structural features, preserve complex boundaries and textures, and simultaneously suppress noise effectively. We establish the existence of a solution for the proposed model and employ an efficient alternating minimization algorithm for numerical solution. Numerical experiments on synthetic, natural, and medical images demonstrate the desirable performance of the proposed model.</div></div>","PeriodicalId":49523,"journal":{"name":"Signal Processing","volume":"241 ","pages":"Article 110410"},"PeriodicalIF":3.6,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145579652","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}

引用次数: 0

Synchrosqueezed windowed linear canonical transform: A method for mode retrieval from multicomponent signals with crossing instantaneous frequencies 同步压缩加窗线性正则变换：一种从具有交叉瞬时频率的多分量信号中提取模式的方法

IF 3.6 2区工程技术 Q2 ENGINEERING, ELECTRICAL & ELECTRONIC

Signal Processing

Pub Date : 2026-04-01 Epub Date: 2025-11-05 DOI: 10.1016/j.sigpro.2025.110384

Shuixin Li , Jiecheng Chen , Qingtang Jiang , Jian Lu

In nature, signals often appear in the form of the superposition of multiple non-stationary signals. The overlap of signal components in the time–frequency domain poses a significant challenge for signal analysis. One approach to addressing this problem is to introduce an additional chirprate parameter and use the chirplet transform (CT) to elevate the two-dimensional time–frequency representation to a three-dimensional time–frequency–chirprate representation. From a certain point of view, the CT of a signal can be regarded as a special windowed linear canonical transform of that signal, undergoing a shift and a modulation.

In this paper, we develop this idea to propose a novel windowed linear canonical transform (WLCT), which provides a new time–frequency–chirprate representation. We discuss four types of WLCTs. In addition, we use a special X-ray transform to further sharpen the time–frequency–chirprate representation. Furthermore, we derive the corresponding three-dimensional synchrosqueezed transform, demonstrating that the WLCTs have great potential for three-dimensional signal separation.

在自然界中，信号往往以多个非平稳信号的叠加形式出现。信号分量在时频域的重叠给信号分析带来了很大的挑战。解决这一问题的一种方法是引入一个额外的chirprate参数，并使用chirplet变换（CT）将二维时频表示提升为三维时频chirprate表示。从某种角度来看，信号的CT可以看作是该信号经过移位和调制后的一个特殊的加窗线性正则变换。在本文中，我们发展了这一思想，提出了一种新的加窗线性正则变换（WLCT），它提供了一种新的时频啁啾表示。我们将讨论四种类型的wlct。此外，我们使用特殊的x射线变换来进一步锐化时频啁啾表示。此外，我们推导了相应的三维同步压缩变换，证明了wlct在三维信号分离方面具有很大的潜力。

{"title":"Synchrosqueezed windowed linear canonical transform: A method for mode retrieval from multicomponent signals with crossing instantaneous frequencies","authors":"Shuixin Li , Jiecheng Chen , Qingtang Jiang , Jian Lu","doi":"10.1016/j.sigpro.2025.110384","DOIUrl":"10.1016/j.sigpro.2025.110384","url":null,"abstract":"<div><div>In nature, signals often appear in the form of the superposition of multiple non-stationary signals. The overlap of signal components in the time–frequency domain poses a significant challenge for signal analysis. One approach to addressing this problem is to introduce an additional chirprate parameter and use the chirplet transform (CT) to elevate the two-dimensional time–frequency representation to a three-dimensional time–frequency–chirprate representation. From a certain point of view, the CT of a signal can be regarded as a special windowed linear canonical transform of that signal, undergoing a shift and a modulation.</div><div>In this paper, we develop this idea to propose a novel windowed linear canonical transform (WLCT), which provides a new time–frequency–chirprate representation. We discuss four types of WLCTs. In addition, we use a special X-ray transform to further sharpen the time–frequency–chirprate representation. Furthermore, we derive the corresponding three-dimensional synchrosqueezed transform, demonstrating that the WLCTs have great potential for three-dimensional signal separation.</div></div>","PeriodicalId":49523,"journal":{"name":"Signal Processing","volume":"241 ","pages":"Article 110384"},"PeriodicalIF":3.6,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145475629","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}

引用次数: 0

Inter-array beam cross-correlation for spatially close multipath discrimination in the reliable acoustic path 可靠声程中空间近距离多径识别的阵列间波束互相关

IF 3.6 2区工程技术 Q2 ENGINEERING, ELECTRICAL & ELECTRONIC

Signal Processing

Pub Date : 2026-04-01 Epub Date: 2025-11-17 DOI: 10.1016/j.sigpro.2025.110411

Ning Wang , Rui Duan , Zhanchao Liu , Xiaoyi Zhou

The directions of arrival (DOA) of the direct path (D) and the surface-reflected path (SR) from a submerged moving source to deep receivers provide range and depth information for source localization in Reliable Acoustic Path (RAP) environments. However, the grazing angles of the D and SR paths are typically very close, posing significant challenges for accurate discrimination. This study proposes a high-resolution multipath DOA estimation method based on horizontally distributed vertical line arrays (VLAs) in the deep sea. This method applies inter-array beam cross-correlation in the time delay–Doppler shift domain, thereby distinguishing the multipath arrivals in this two-dimensional domain rather than the one-dimensional domain of angle. The process also benefits from the spatial gain of beamforming and the temporal (pulse compression) gain of inter-beam cross-correlation, yielding two key advantages: (1) substantial enhancement of the signal-to-noise ratio (SNR), facilitating the reliable extraction of weak multipath arrivals, particularly the SR; and (2) improved spatial filtering, which effectively suppresses uncorrelated interference between arrays. Experimental results from the South China Sea demonstrate that the proposed method accurately extracts the DOAs of both the D and SR paths of a moving source at a depth of approximately 72 meters, even under significant surface ship interference. In contrast, conventional single-array high-resolution methods, such as compressive sensing and MVDR, fail to extract the SR path.

在可靠声路径（RAP）环境中，水下移动源到深部接收器的直接路径(D)和表面反射路径（SR）的到达方向（DOA）为源定位提供了范围和深度信息。然而，D和SR路径的掠角通常非常接近，这对准确识别构成了重大挑战。提出了一种基于水平分布垂直线阵列（VLAs）的深海高分辨率多径DOA估计方法。该方法在时延-多普勒频移域应用阵列间波束互相关，从而在该二维域而不是一维角度域中区分多径到达。该过程还受益于波束形成的空间增益和波束间互相关的时间（脉冲压缩）增益，产生两个关键优势：(1)显著提高了信噪比（SNR），有助于可靠地提取弱多径到达，特别是SR；(2)改进空间滤波，有效抑制阵列间不相关干扰。南海实验结果表明，该方法在水面舰艇明显干扰的情况下，仍能准确提取深度约为72米的动源D路径和SR路径的doa。相比之下，传统的单阵列高分辨率方法，如压缩感知和MVDR，无法提取SR路径。

{"title":"Inter-array beam cross-correlation for spatially close multipath discrimination in the reliable acoustic path","authors":"Ning Wang , Rui Duan , Zhanchao Liu , Xiaoyi Zhou","doi":"10.1016/j.sigpro.2025.110411","DOIUrl":"10.1016/j.sigpro.2025.110411","url":null,"abstract":"<div><div>The directions of arrival (DOA) of the direct path (D) and the surface-reflected path (SR) from a submerged moving source to deep receivers provide range and depth information for source localization in Reliable Acoustic Path (RAP) environments. However, the grazing angles of the D and SR paths are typically very close, posing significant challenges for accurate discrimination. This study proposes a high-resolution multipath DOA estimation method based on horizontally distributed vertical line arrays (VLAs) in the deep sea. This method applies inter-array beam cross-correlation in the time delay<strong>–</strong>Doppler shift domain, thereby distinguishing the multipath arrivals in this two-dimensional domain rather than the one-dimensional domain of angle. The process also benefits from the spatial gain of beamforming and the temporal (pulse compression) gain of inter-beam cross-correlation, yielding two key advantages: (1) substantial enhancement of the signal-to-noise ratio (SNR), facilitating the reliable extraction of weak multipath arrivals, particularly the SR; and (2) improved spatial filtering, which effectively suppresses uncorrelated interference between arrays. Experimental results from the South China Sea demonstrate that the proposed method accurately extracts the DOAs of both the D and SR paths of a moving source at a depth of approximately 72 meters, even under significant surface ship interference. In contrast, conventional single-array high-resolution methods, such as compressive sensing and MVDR, fail to extract the SR path.</div></div>","PeriodicalId":49523,"journal":{"name":"Signal Processing","volume":"241 ","pages":"Article 110411"},"PeriodicalIF":3.6,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145624470","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}

引用次数: 0

Recursive filter for 2-D Markov jump systems under multi-channel deception attacks and weighted try-once-discard protocol 二维马尔可夫跳变系统在多信道欺骗攻击下的递归滤波和加权尝试一次丢弃协议

IF 3.6 2区工程技术 Q2 ENGINEERING, ELECTRICAL & ELECTRONIC

Signal Processing

Pub Date : 2026-04-01 Epub Date: 2025-11-08 DOI: 10.1016/j.sigpro.2025.110391

Yufan Wang, Chunyan Han

This paper investigates the state estimation for 2-D Markov jump systems with stochastic nonlinearity, and multichannel deception attacks under the weighted try-once-discard (WTOD) scheduling protocol. A novel multi-channel deception attack model is developed within a 2-D framework, in which the deception attacks are assumed to occur in two classes of channels: the output channels between the plant and sensors and the communication channels from the sensors to the estimator. The status of the multi-channel attacks is governed by two diagonal matrices, where each diagonal element is modeled as a Bernoulli random variable, and the attack signals are energy bounded. Furthermore, a 2-D WTOD protocol is developed to schedule data transmission, which orchestrates the nodes to access the network based on a quadratic selection principle. The main aim of this paper is to design a recursive estimator that minimizes the upper bounds (UBs) of the error variances (EVs) in the presence of two-channel attacks and the WTOD protocol. By means of mathematical induction and matrix inequality techniques, certain UBs are attained on the EVs. The filter gain is then computed at each step to minimize the devised UBs by solving a set of 2-D Riccati difference equations.

研究了加权尝试一次丢弃调度协议下二维随机非线性马尔可夫跳变系统的状态估计和多信道欺骗攻击问题。在二维框架内建立了一种新的多通道欺骗攻击模型，该模型假设欺骗攻击发生在两类通道上：设备与传感器之间的输出通道和传感器与估计器之间的通信通道。多通道攻击的状态由两个对角矩阵控制，其中每个对角元素被建模为一个伯努利随机变量，攻击信号是能量有界的。在此基础上，提出了一种二维WTOD协议来调度数据传输，该协议基于二次选择原则编排节点访问网络。本文的主要目的是设计一个递归估计器，在存在双通道攻击和WTOD协议的情况下最小化误差方差（ev）的上界（UBs）。利用数学归纳法和矩阵不等式技术，在ev上得到了一定的UBs。然后在每一步计算滤波器增益，通过求解一组二维里卡蒂差分方程来最小化所设计的UBs。

{"title":"Recursive filter for 2-D Markov jump systems under multi-channel deception attacks and weighted try-once-discard protocol","authors":"Yufan Wang, Chunyan Han","doi":"10.1016/j.sigpro.2025.110391","DOIUrl":"10.1016/j.sigpro.2025.110391","url":null,"abstract":"<div><div>This paper investigates the state estimation for 2-D Markov jump systems with stochastic nonlinearity, and multichannel deception attacks under the weighted try-once-discard (WTOD) scheduling protocol. A novel multi-channel deception attack model is developed within a 2-D framework, in which the deception attacks are assumed to occur in two classes of channels: the output channels between the plant and sensors and the communication channels from the sensors to the estimator. The status of the multi-channel attacks is governed by two diagonal matrices, where each diagonal element is modeled as a Bernoulli random variable, and the attack signals are energy bounded. Furthermore, a 2-D WTOD protocol is developed to schedule data transmission, which orchestrates the nodes to access the network based on a quadratic selection principle. The main aim of this paper is to design a recursive estimator that minimizes the upper bounds (UBs) of the error variances (EVs) in the presence of two-channel attacks and the WTOD protocol. By means of mathematical induction and matrix inequality techniques, certain UBs are attained on the EVs. The filter gain is then computed at each step to minimize the devised UBs by solving a set of 2-D Riccati difference equations.</div></div>","PeriodicalId":49523,"journal":{"name":"Signal Processing","volume":"241 ","pages":"Article 110391"},"PeriodicalIF":3.6,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145529035","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}

引用次数: 0

Fast and efficient implementation of the maximum likelihood estimation for the linear regression with Gaussian model uncertainty 具有高斯模型不确定性的线性回归的最大似然估计的快速有效实现

IF 3.6 2区工程技术 Q2 ENGINEERING, ELECTRICAL & ELECTRONIC

Signal Processing

Pub Date : 2026-04-01 Epub Date: 2025-11-12 DOI: 10.1016/j.sigpro.2025.110403

Ruohai Guo , Jiang Zhu , Xing Jiang , Fengzhong Qu

The linear regression model with a random variable (RV) measurement matrix, where the mean of the random measurement matrix has full column rank, has been extensively studied. In particular, the quasiconvexity of the maximum likelihood estimation (MLE) problem was established, and the corresponding Cramér–Rao bound (CRB) was derived, leading to the development of an efficient bisection-based algorithm known as RV-ML. In contrast, this work extends the analysis to both overdetermined and underdetermined cases, allowing the mean of the random measurement matrix to be rank-deficient. A remarkable contribution is the proof that the equivalent MLE problem is convex and satisfies strong duality, strengthening previous quasiconvexity results. Moreover, it is shown that in underdetermined scenarios, the randomness in the measurement matrix can be beneficial for estimation under certain conditions. In addition, a fast and unified implementation of the MLE solution, referred to as generalized RV-ML (GRV-ML), is proposed, which handles a more general case including both underdetermined and overdetermined systems. Extensive numerical simulations are provided to validate the theoretical findings.

随机测量矩阵均值为全列秩的随机变量测量矩阵线性回归模型得到了广泛的研究。特别地，建立了极大似然估计（MLE）问题的拟自凸性，并推导了相应的cram - rao界（CRB），从而发展了一种高效的基于二分法的RV-ML算法。相反，这项工作将分析扩展到过确定和欠确定的情况，允许随机测量矩阵的平均值是秩不足的。一个值得注意的贡献是证明了等效MLE问题是凸的并且满足强对偶性，加强了以前的拟凸性结果。此外，在欠确定的情况下，测量矩阵的随机性在一定条件下有利于估计。此外，还提出了一种快速统一的MLE解决方案，称为广义RV-ML (GRV-ML)，它可以处理更一般的情况，包括欠定和过定系统。提供了大量的数值模拟来验证理论结果。

{"title":"Fast and efficient implementation of the maximum likelihood estimation for the linear regression with Gaussian model uncertainty","authors":"Ruohai Guo , Jiang Zhu , Xing Jiang , Fengzhong Qu","doi":"10.1016/j.sigpro.2025.110403","DOIUrl":"10.1016/j.sigpro.2025.110403","url":null,"abstract":"<div><div>The linear regression model with a random variable (RV) measurement matrix, where the mean of the random measurement matrix has full column rank, has been extensively studied. In particular, the quasiconvexity of the maximum likelihood estimation (MLE) problem was established, and the corresponding Cramér–Rao bound (CRB) was derived, leading to the development of an efficient bisection-based algorithm known as RV-ML. In contrast, this work extends the analysis to both overdetermined and underdetermined cases, allowing the mean of the random measurement matrix to be rank-deficient. A remarkable contribution is the proof that the equivalent MLE problem is convex and satisfies strong duality, strengthening previous quasiconvexity results. Moreover, it is shown that in underdetermined scenarios, the randomness in the measurement matrix can be beneficial for estimation under certain conditions. In addition, a fast and unified implementation of the MLE solution, referred to as generalized RV-ML (GRV-ML), is proposed, which handles a more general case including both underdetermined and overdetermined systems. Extensive numerical simulations are provided to validate the theoretical findings.</div></div>","PeriodicalId":49523,"journal":{"name":"Signal Processing","volume":"241 ","pages":"Article 110403"},"PeriodicalIF":3.6,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145579653","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}

引用次数: 0

A regular superpixel generation method based on continuous edges 一种基于连续边缘的规则超像素生成方法

IF 3.6 2区工程技术 Q2 ENGINEERING, ELECTRICAL & ELECTRONIC

Signal Processing

Pub Date : 2026-04-01 Epub Date: 2025-11-21 DOI: 10.1016/j.sigpro.2025.110412

Daipeng Yang , Bo Peng , Tingyu Zhao , Xiaofan Li

Regular superpixels are often preferred because they can be more easily applied to subsequent tasks. Existing superpixel boundaries are often highly irregular, frequently exhibiting jagged and rough edges. In the human visual system, the fourth visual cortex is typically considered the region responsible for image segmentation. Edge information from lower visual cortices is integrated in the fourth visual cortex and used to highlight the foreground and facilitate image segmentation. Inspired by this, we propose a regular superpixel generation method based on continuous edges. Our method first models orientation-selective neurons and use their orientation properties to thin and connect edges, resulting in continuous edge segments. These edge segments are then extended and reasonably partitioned by regular grid edges to form closed regions. Finally, we extract the closed regions and merge them to generate the desired number of superpixels. Experiments show that, compared to other superpixel generation methods, the superpixels obtained by our method are more regular. In smooth regions, our method produces square-shaped superpixels, while in non-smooth regions, the boundaries of our superpixels either closedly adhere to the object contours or align with the grid edges, effectively segmenting the object regions. The source code for our method is available at https://github.com/DaipengYang7/RSCE.

常规超像素通常是首选，因为它们可以更容易地应用于后续任务。现有的超像素边界通常非常不规则，经常表现出锯齿状和粗糙的边缘。在人类视觉系统中，第四视觉皮层通常被认为是负责图像分割的区域。来自下视觉皮层的边缘信息被整合到第四视觉皮层，用于突出前景和促进图像分割。受此启发，我们提出了一种基于连续边缘的规则超像素生成方法。我们的方法首先对定向选择神经元进行建模，并利用它们的定向属性来细化和连接边缘，从而得到连续的边缘段。然后将这些边段用规则的网格边进行扩展和合理分割，形成封闭区域。最后，我们提取封闭区域并合并它们以生成所需的超像素数量。实验表明，与其他超像素生成方法相比，本文方法得到的超像素更有规则性。在光滑区域，我们的方法产生方形的超像素，而在非光滑区域，我们的超像素的边界要么紧密地附着在物体轮廓上，要么与网格边缘对齐，有效地分割了物体区域。我们的方法的源代码可从https://github.com/DaipengYang7/RSCE获得。

{"title":"A regular superpixel generation method based on continuous edges","authors":"Daipeng Yang , Bo Peng , Tingyu Zhao , Xiaofan Li","doi":"10.1016/j.sigpro.2025.110412","DOIUrl":"10.1016/j.sigpro.2025.110412","url":null,"abstract":"<div><div>Regular superpixels are often preferred because they can be more easily applied to subsequent tasks. Existing superpixel boundaries are often highly irregular, frequently exhibiting jagged and rough edges. In the human visual system, the fourth visual cortex is typically considered the region responsible for image segmentation. Edge information from lower visual cortices is integrated in the fourth visual cortex and used to highlight the foreground and facilitate image segmentation. Inspired by this, we propose a regular superpixel generation method based on continuous edges. Our method first models orientation-selective neurons and use their orientation properties to thin and connect edges, resulting in continuous edge segments. These edge segments are then extended and reasonably partitioned by regular grid edges to form closed regions. Finally, we extract the closed regions and merge them to generate the desired number of superpixels. Experiments show that, compared to other superpixel generation methods, the superpixels obtained by our method are more regular. In smooth regions, our method produces square-shaped superpixels, while in non-smooth regions, the boundaries of our superpixels either closedly adhere to the object contours or align with the grid edges, effectively segmenting the object regions. The source code for our method is available at <span><span>https://github.com/DaipengYang7/RSCE</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":49523,"journal":{"name":"Signal Processing","volume":"241 ","pages":"Article 110412"},"PeriodicalIF":3.6,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145624469","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}

引用次数: 0