IEEE Journal of Selected Topics in Signal Processing最新文献

{"title":"CoMCM:Collaborative 3D Detection With Multiscale Clustering Mamba","authors":"Tong Wang;Jie Guo;Ming Ouyang;Peng Xue;Lu Wang;Pei Xiao","doi":"10.1109/JSTSP.2025.3650028","DOIUrl":"https://doi.org/10.1109/JSTSP.2025.3650028","url":null,"abstract":"Collaborative multi-agent sensing via Vehicle-to-Vehicle (V2V) and Vehicle-to-Everything (V2X) communication has emerged as a promising solution to overcome the limitations of single-vehicle perception in autonomous driving. However, efficiently fusing large-scale, high-dimensional features across multiple vehicles remains a significant challenge, particularly under communication localization constraints. This paper proposes a collaborative 3D object detection framework, termed Collaborative 3D Detection with Multiscale Clustering Mamba (CoMCM). It performs adaptive feature fusion across multiple spatial scales, integrating both coarse- and fine-grained information. CoMCM comprises two core components: the Contextual Clustering Mamba (CCMamba) and a collaborative Mamba in BEV fusion module(CoM). The MCMamba module incorporates multiscale clustering within a Mamba-based state-space model to capture both global and local contextual information. This design addresses the limitations of selective state modeling in non-causal BEV representations. The CoM module further fuses BEV features from multiple Connected and Automated Vehicles (CAVs) using relative pose-aware attention and adaptive weighting, enabling effective multi-vehicle collaboration. Extensive experiments on large-scale datasets, OPV2V, V2XSet and the real-world DAIR-V2X datasets, demonstrate that CoMCM significantly outperforms existing collaborative 3D object detection methods and remains robust under bandwidth limitations and pose estimation errors. Moreover, CoMCM achieves low computational cost while maintaining high detection accuracy. CoMCM lays the foundation for scalable and accurate collaborative perception in intelligent connected vehicle systems operating in complex environments.","PeriodicalId":13038,"journal":{"name":"IEEE Journal of Selected Topics in Signal Processing","volume":"19 8","pages":"1996-2009"},"PeriodicalIF":13.7,"publicationDate":"2025-12-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146122798","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"IEEE Signal Processing Society Information","authors":"","doi":"10.1109/JSTSP.2025.3647358","DOIUrl":"https://doi.org/10.1109/JSTSP.2025.3647358","url":null,"abstract":"","PeriodicalId":13038,"journal":{"name":"IEEE Journal of Selected Topics in Signal Processing","volume":"19 7","pages":"C3-C3"},"PeriodicalIF":13.7,"publicationDate":"2025-12-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11320991","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145861203","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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":"IEEE Signal Processing Society Information","authors":"","doi":"10.1109/JSTSP.2025.3644519","DOIUrl":"https://doi.org/10.1109/JSTSP.2025.3644519","url":null,"abstract":"","PeriodicalId":13038,"journal":{"name":"IEEE Journal of Selected Topics in Signal Processing","volume":"19 6","pages":"C3-C3"},"PeriodicalIF":13.7,"publicationDate":"2025-12-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11319270","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145852465","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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":"FluoEcoli-Instance: A High-Content Fluorescence Microscopy Dataset and CBAM-Enhanced Hybrid Task Cascade for E. Coli Instance Segmentation","authors":"Son T. Huynh;Hai X. Nguyen;Dat Q. Duong;Tien K. Nguyen;Nhi K. P. Nguyen;Hoang Tran Le Nguyen;Phuong N. Vu;Nghi P. G. Nguyen;Khoi D. Vu;Hoa Q. N. Ho;Tuan-Anh Tran;Stephen Baker;Binh T. Nguyen","doi":"10.1109/JSTSP.2025.3648286","DOIUrl":"https://doi.org/10.1109/JSTSP.2025.3648286","url":null,"abstract":"Antimicrobial resistance (AMR) has become a global health crisis, creating an urgent need for rapid and accurate bacterial identification to guide appropriate antibiotic therapy. Automated segmentation of bacteria in microscopic images enables quantitative analysis of cellular morphology and spatial patterns, offering valuable cues for early species recognition. Such image-based insights can accelerate diagnostic decisions and promote rational antibiotic use, ultimately mitigating the impact of AMR. Hence, we present a significantly expanded dataset of <italic>E. coli</i> fluorescence images, extending that of Dat et al. (Duong et al., 2023) with additional samples and refined annotations to better support bacterial segmentation research. We extensively re-examine several popular segmentation models using standardised COCO metrics <bold>(<inline-formula><tex-math>$mAP$</tex-math></inline-formula>, <inline-formula><tex-math>$mAP_{0.50}$</tex-math></inline-formula>, <inline-formula><tex-math>$mAP_{0.75}$</tex-math></inline-formula>, <inline-formula><tex-math>$mAP_{text{small}}$</tex-math></inline-formula>)</b> and find that methods originally designed for semantic segmentation struggling with instance-level tasks. On the other hand, two-stage detectors, especially Cascade Mask R-CNN (Vasconcelos et al., 2018) and its Hybrid Task Cascade (HTC) (Chen et al., 2019) variant, perform best when paired with modern backbone networks. In particular, our framework introduces a novel synergy between the ConvNeXtV2-Tiny backbone and the HTC architecture, enhanced by the CBAM (Kweon et al., 2018) attention mechanism (reduction ratio of eight), to achieve superior feature refinement and instance delineation. This configuration sets a new performance benchmark, achieving <inline-formula><tex-math>$mAP_{0.50} = 0.940$</tex-math></inline-formula>, <inline-formula><tex-math>$mAP_{0.75} = 0.882$</tex-math></inline-formula>, <inline-formula><tex-math>$mAP_{text{small}} = 0.788$</tex-math></inline-formula>, and <inline-formula><tex-math>$mAP_{0.50:0.95} = 0.787$</tex-math></inline-formula>, and directly supports the development of data-driven, clinically relevant microbial analysis systems that can guide evidence-based and timely antimicrobial treatment strategies.","PeriodicalId":13038,"journal":{"name":"IEEE Journal of Selected Topics in Signal Processing","volume":"19 8","pages":"1967-1982"},"PeriodicalIF":13.7,"publicationDate":"2025-12-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146122732","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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 Survey on Speech Large Language Models for Understanding","authors":"Jing Peng;Yucheng Wang;Bohan Li;Yiwei Guo;Hankun Wang;YanGui Fang;Yu Xi;Haoyu Li;Xu Li;Ke Zhang;Shuai Wang;Kai Yu","doi":"10.1109/JSTSP.2025.3640535","DOIUrl":"https://doi.org/10.1109/JSTSP.2025.3640535","url":null,"abstract":"Speech understanding is essential for interpreting the diverse forms of information embedded in spoken language, including linguistic, paralinguistic, and non-linguistic cues that are vital for effective human-computer interaction. The rapid advancement of large language models (LLMs) has catalyzed the emergence of Speech Large Language Models (Speech LLMs), which marks a transformative shift toward general-purpose speech understanding systems. To further clarify and systematically delineate task objectives, in this paper, we formally define the concept of speech understanding and introduce a structured taxonomy encompassing its informational, functional, and format dimensions. Within this scope of definition, we present a comprehensive review of current Speech LLMs, analyzing their architectures through a three-stage abstraction: Modality Feature Extraction, Modality Information Fusion, and LLM Inference. In addition, we examine training strategies, discuss representative datasets, and review evaluation methodologies adopted in the field. Based on empirical analyses and experimental evidence, we identify two key challenges currently facing Speech LLMs—instruction sensitivity and degradation in semantic reasoning—and propose concrete directions for addressing these issues. Through this systematic and detailed survey, we aim to offer a foundational reference for researchers and practitioners working toward more robust, generalizable, and human-aligned Speech LLMs.","PeriodicalId":13038,"journal":{"name":"IEEE Journal of Selected Topics in Signal Processing","volume":"20 1","pages":"2-31"},"PeriodicalIF":13.7,"publicationDate":"2025-12-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147274990","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Dynamic Array-of-Subarrays Architecture for Wideband Multi-Antenna Systems","authors":"Seungnyun Kim;Moe Z. Win","doi":"10.1109/JSTSP.2025.3539208","DOIUrl":"https://doi.org/10.1109/JSTSP.2025.3539208","url":null,"abstract":"Recently, wideband beamforming realized by extremely large-scale antenna array (ELAA) systems have garnered significant interest as a means to dramatically improve throughput of next generation (xG) networks. However, traditional phase shifter (PS)-based beamforming schemes might not work well in wideband ELAA systems due to the beam squint effect, where beams at different frequencies become misaligned. While the use of true time delay (TTD) can mitigate the beam squint effect by generating frequency-dependent beamforming vectors, the conventional TTD-based beamforming schemes suffer from significant array gain loss caused by the discrepancies between the desired directional beams and the generated beams. This beam misalignment issue becomes even more pronounced in the wideband ELAA systems due to the nonlinear near-field characteristics described by the spherical wave propagation. In this paper, we propose a wideband dynamic array-of-subarrays (WDAoSA) architecture that dynamically configures connections between TTDs and PS subarrays using a switch network. By optimizing the subarray connections as well as the TTD time delays and PS phase shifts, WDAoSA can effectively maximize the array gain of wideband ELAA systems. Numerical results demonstrate that WDAoSA achieves significant improvements in terms of array gain and spectral efficiency over the conventional TTD-based beamforming schemes.","PeriodicalId":13038,"journal":{"name":"IEEE Journal of Selected Topics in Signal Processing","volume":"19 5","pages":"840-855"},"PeriodicalIF":13.7,"publicationDate":"2025-12-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145659230","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"IEEE Signal Processing Society Publication Information","authors":"","doi":"10.1109/JSTSP.2025.3607338","DOIUrl":"https://doi.org/10.1109/JSTSP.2025.3607338","url":null,"abstract":"","PeriodicalId":13038,"journal":{"name":"IEEE Journal of Selected Topics in Signal Processing","volume":"19 5","pages":"C2-C2"},"PeriodicalIF":13.7,"publicationDate":"2025-12-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11275883","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145659234","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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":"IEEE Signal Processing Society Information","authors":"","doi":"10.1109/JSTSP.2025.3607336","DOIUrl":"https://doi.org/10.1109/JSTSP.2025.3607336","url":null,"abstract":"","PeriodicalId":13038,"journal":{"name":"IEEE Journal of Selected Topics in Signal Processing","volume":"19 5","pages":"C3-C3"},"PeriodicalIF":13.7,"publicationDate":"2025-12-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11275990","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145659187","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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":"Spatial Invertible Network With Mamba-Convolution for Hyperspectral Image Fusion","authors":"Lin Xiao;Shuo Guo;Feng Mo;Qiya Song;Yining Yang;Yuanye Liu;Xiaohui Wei;Tao Yang;Renwei Dian","doi":"10.1109/JSTSP.2025.3596518","DOIUrl":"10.1109/JSTSP.2025.3596518","url":null,"abstract":"Due to the complementary characteristics of hyperspectral image (HSI) and multispectral image (MSI), fusing them can generate a high-resolution hyperspectral image (HR-HSI). However, existing convolutional neural network (CNN)- and Transformer-based methods still suffer from challenges such as spectral distortion, insufficient texture representation, and limited ability to integrate global-local information. To address these issues, we propose a spatial invertible network (SINet) with Mamba-convolution. First, we design the multiscale invertible spatial (MISpa) module to effectively model the spatial information of the source image. The MISpa module utilizes Haar wavelet decomposition to separate high and low-frequency spatial components. By integrating invertible neural network (INN), it enhances the extraction of discriminative spatial features while maintaining structural integrity. To mitigate the risk of excessive reliance on spatial information, which may lead to spectral distortion, we employ spectral compression (via average pooling and 3 × 3 convolution) to extract spectral information at each scale of the source image. Through this way, SINet establishes an effective trade-off between detail enhancement and spectral fidelity. Second, we propose a Mamba convolutional fusion module (MCFM) that integrates feature extraction capabilities of Mamba and CNN. This module is designed to adaptively enhance the complementarity between global and local features, thereby guiding the process of feature enhancement. Experiments on both simulated and real-world datasets from quantitative and qualitative perspectives to validate the superiority of the proposed method compared to existing approaches.","PeriodicalId":13038,"journal":{"name":"IEEE Journal of Selected Topics in Signal Processing","volume":"19 8","pages":"1643-1653"},"PeriodicalIF":8.5,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147879211","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"DUN-SRE: Deep Unrolling Network With Spatiotemporal Rotation Equivariance for Dynamic MRI Reconstruction","authors":"Yuliang Zhu;Jing Cheng;Qi Xie;Zhuo-Xu Cui;Qingyong Zhu;Yuanyuan Liu;Xin Liu;Jianfeng Ren;Chengbo Wang;Dong Liang","doi":"10.1109/JSTSP.2025.3618589","DOIUrl":"https://doi.org/10.1109/JSTSP.2025.3618589","url":null,"abstract":"Dynamic Magnetic Resonance Imaging (MRI) exhibits transformation symmetries, including spatial rotation symmetry within individual frames and temporal symmetry along the time dimension. Explicit incorporation of these symmetry priors in the reconstruction model can significantly improve image quality, especially under aggressive undersampling scenarios. Recently, Equivariant convolutional neural network (ECNN) has shown great promise in exploiting spatial symmetry priors. However, existing ECNNs critically fail to model temporal symmetry, arguably the most universal and informative structural prior in dynamic MRI reconstruction. To tackle this issue, we propose a novel Deep Unrolling Network with Spatiotemporal Rotation Equivariance (DUN-SRE) for Dynamic MRI Reconstruction. The DUN-SRE establishes spatiotemporal equivariance through a (2+1)D equivariant convolutional architecture. In particular, it integrates both the data consistency and proximal mapping module into a unified deep unrolling framework. This architecture ensures rigorous propagation of spatiotemporal rotation symmetry constraints throughout the reconstruction process, enabling more physically accurate modeling of cardiac motion dynamics in cine MRI. In addition, a high-fidelity group filter parameterization mechanism is developed to maintain representation precision while enforcing symmetry constraints. Comprehensive experiments on Cardiac CINE MRI datasets demonstrate that DUN-SRE achieves state-of-the-art performance, particularly in preserving rotation-symmetric structures, offering strong generalization capability to a broad range of dynamic MRI reconstruction tasks.","PeriodicalId":13038,"journal":{"name":"IEEE Journal of Selected Topics in Signal Processing","volume":"19 8","pages":"1694-1709"},"PeriodicalIF":8.5,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148504673","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}