IEEE Signal Processing Letters最新文献_第2页

MMA: Video Reconstruction for Spike Camera Based on Multiscale Temporal Modeling and Fine-Grained Attention

IF 3.2 2区工程技术

IEEE Signal Processing Letters Pub Date : 2025-03-11 DOI: 10.1109/LSP.2025.3550278

Dilmurat Alim;Chen Yang;Laiyun Qing;Guorong Li;Qingming Huang

{"title":"MMA: Video Reconstruction for Spike Camera Based on Multiscale Temporal Modeling and Fine-Grained Attention","authors":"Dilmurat Alim;Chen Yang;Laiyun Qing;Guorong Li;Qingming Huang","doi":"10.1109/LSP.2025.3550278","DOIUrl":"https://doi.org/10.1109/LSP.2025.3550278","url":null,"abstract":"This paper presents a Multiscale Temporal Correlation Learning with the Mamba-Fused Attention Model (MMA), an efficient and effective method for reconstructing a video clip from a spike stream. Spike cameras offer unique advantages for capturing rapid scene changes with high temporal resolution. A spike stream contains sufficient information for multiple image reconstructions. However, existing methods generate only a single image at a time for a given spike stream, which results in excessive redundant computations between consecutive frames when aiming at restoring a video clip, thereby increasing computational costs significantly. The proposed MMA addresses such challenges by constructing a spike-to-video model, directly producing an image sequence at a time. Specifically, we propose a U-shaped Multiscale Temporal Correlation Learning (MTCL) to fuse the features at different temporal resolutions for clear video reconstruction. At each scale, we introduce a Fine-Grained Attention (FGA) module for fine-spatial context modeling within a patch and a Mamba module for integrating features across patches. Adopting a lightweight U-shaped structure and fine-grained feature extraction at each level, our method reconstructs high-quality image sequences quickly. The experimental results show that the proposed MMA surpasses current state-of-the-art methods in image quality, computation cost, and model size.","PeriodicalId":13154,"journal":{"name":"IEEE Signal Processing Letters","volume":"32 ","pages":"1291-1295"},"PeriodicalIF":3.2,"publicationDate":"2025-03-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143716508","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

Vision-Inspired Boundary Perception Network for Lightweight Camouflaged Object Detection

IF 3.2 2区工程技术

IEEE Signal Processing Letters Pub Date : 2025-03-11 DOI: 10.1109/LSP.2025.3549698

Chunyuan Chen;Weiyun Liang;Donglin Wang;Bin Wang;Jing Xu

{"title":"Vision-Inspired Boundary Perception Network for Lightweight Camouflaged Object Detection","authors":"Chunyuan Chen;Weiyun Liang;Donglin Wang;Bin Wang;Jing Xu","doi":"10.1109/LSP.2025.3549698","DOIUrl":"https://doi.org/10.1109/LSP.2025.3549698","url":null,"abstract":"Lightweight camouflaged object detection (COD) has garnered increasing attention due to its wide range of real-world applications and its efficiency on mobile devices. Existing lightweight COD methods typically attempt to utilize multi-scale fusion, frequency cues, and texture information to enhance the representation ability of lightweight backbone features. However, they still fall short in detecting precise and continuous object boundaries. To address this issue, we observe that two types of cells in the human visual system make great contributions to boundary perception. Motivated by this, we propose a boundary perception module (BPM) to enhance features with the awareness of fine-grained boundary, by mimicking the boundary perception process of aforementioned cells. In addition, we propose a bidirectional semantic enhancement module (BSEM) to effectively decode multi-level features in a lightweight manner. With BPM and BSEM, our proposed vision-inspired boundary perception network (BPNet) achieves superior performance against state-of-the-art methods and surpasses lightweight COD models by a large margin with the least parameters (3.64 M) and fastest speed (168FPS for the input size of 384 × 384).","PeriodicalId":13154,"journal":{"name":"IEEE Signal Processing Letters","volume":"32 ","pages":"1176-1180"},"PeriodicalIF":3.2,"publicationDate":"2025-03-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143667270","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

Dual-Stream Image Sharing Chain Detection via Dynamic Information Compensation

IF 3.2 2区工程技术

IEEE Signal Processing Letters Pub Date : 2025-03-11 DOI: 10.1109/LSP.2025.3550282

Xinyi Su;Yuanman Li;Yulong Zheng;Xia Li

引用次数: 0

Power Transformed Density Ridge Estimation

IF 3.2 2区工程技术

IEEE Signal Processing Letters Pub Date : 2025-03-10 DOI: 10.1109/LSP.2025.3549699

Hengchao Chen;Zheng Zhai

{"title":"Power Transformed Density Ridge Estimation","authors":"Hengchao Chen;Zheng Zhai","doi":"10.1109/LSP.2025.3549699","DOIUrl":"https://doi.org/10.1109/LSP.2025.3549699","url":null,"abstract":"This paper proposes to control ridge estimation with nonlinear transformations. We establish an inclusion relationship between ridges with/without transformations: <inline-formula><tex-math>${mathcal R}(fcirc p)subseteq {mathcal R}(p)$</tex-math></inline-formula>, where <inline-formula><tex-math>${mathcal R}(p)$</tex-math></inline-formula> is the <inline-formula><tex-math>$d$</tex-math></inline-formula>-dimensional ridge of <inline-formula><tex-math>$p:mathbb {R}^{D}to mathbb {R}$</tex-math></inline-formula> and <inline-formula><tex-math>$f$</tex-math></inline-formula> is a strictly increasing and concave map defined on range<inline-formula><tex-math>$(p)$</tex-math></inline-formula>. This finding reveals the benefit of applying an increasing and concave transformation before ridge estimation. As <inline-formula><tex-math>$f^{q}(y)=y^{q}/q, qleq 1$</tex-math></inline-formula> are increasing and concave on <inline-formula><tex-math>$mathbb {R}_+$</tex-math></inline-formula>, we use these power transformations to positive density functions, and then perform ridge estimation. The algorithm, named power-transformed subspace-constrained mean-shift (PSCMS), outperforms its competitors in numerical experiments.","PeriodicalId":13154,"journal":{"name":"IEEE Signal Processing Letters","volume":"32 ","pages":"1286-1290"},"PeriodicalIF":3.2,"publicationDate":"2025-03-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143716488","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

Ambiguity-Free Broadband DOA Estimation Relying on Parameterized Time-Frequency Transform

IF 3.2 2区工程技术

IEEE Signal Processing Letters Pub Date : 2025-03-10 DOI: 10.1109/LSP.2025.3550002

Wei Wang;Shefeng Yan;Linlin Mao;Zeping Sui;Jirui Yang

引用次数: 0

Adaptive Alignment and Time Aggregation Network for Speech-Visual Emotion Recognition

IF 3.2 2区工程技术

IEEE Signal Processing Letters Pub Date : 2025-03-10 DOI: 10.1109/LSP.2025.3550007

Lile Wu;Lei Bai;Wenhao Cheng;Zutian Cheng;Guanghui Chen

引用次数: 0

Contextual Direct Position Determination for Path Loss Informed Localization

IF 3.2 2区工程技术

IEEE Signal Processing Letters Pub Date : 2025-03-10 DOI: 10.1109/LSP.2025.3550047

Licheng Zhao;Wenqiang Pu;Rui Zhou;Ming-Yi You;Qingjiang Shi

引用次数: 0

Efficient Feature Focus Enhanced Network for Small and Dense Object Detection in SAR Images

IF 3.2 2区工程技术

IEEE Signal Processing Letters Pub Date : 2025-03-07 DOI: 10.1109/LSP.2025.3548934

Cong Li;Lihu Xi;Yongqiang Hei;Wentao Li;Zhu Xiao

引用次数: 0

Learning Source-Free Domain Adaptation for Infrared Small Target Detection

IF 3.2 2区工程技术

IEEE Signal Processing Letters Pub Date : 2025-03-07 DOI: 10.1109/LSP.2025.3549000

Hongxu Jin;Baiyang Chen;Qianwen Lu;Qingchuan Tao;Yongxiang Li

{"title":"Learning Source-Free Domain Adaptation for Infrared Small Target Detection","authors":"Hongxu Jin;Baiyang Chen;Qianwen Lu;Qingchuan Tao;Yongxiang Li","doi":"10.1109/LSP.2025.3549000","DOIUrl":"https://doi.org/10.1109/LSP.2025.3549000","url":null,"abstract":"Existing infrared small target detection (IRSTD) methods mainly rely on the assumption that the training and testing data come from the same distribution, a premise that does not hold in many real-world scenarios. Additionally, the inability to access source domain data in numerous IRSTD tasks further complicates the domain adaptation process. To address these challenges, we propose a novel Source-Free Domain Adaptation (SFDA) framework for IRSTD, denoted as IRSTD-SFDA. This framework comprises two key components: Multi-expert Domain Adaptation (MDA) and Multi-scale Focused Learning (MFL). MDA leverages the source model to generate pseudo masks for the target domain, facilitating the transfer of knowledge from the source to the target domain. To account for the inherent diversity of small targets across domains, MDA refines these pseudo masks through a series of operations, including target localization, rolling guidance filtering, shape adaptation, and multi-expert decision, thereby mitigating morphological discrepancies between the source and target domains. Meanwhile, MFL employs a global-local fusion strategy to focus on critical regions, enhancing the model's ability to detect small infrared targets. Extensive experimental evaluations across various cross-domain scenarios demonstrate the effectiveness of the proposed framework.","PeriodicalId":13154,"journal":{"name":"IEEE Signal Processing Letters","volume":"32 ","pages":"1121-1125"},"PeriodicalIF":3.2,"publicationDate":"2025-03-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143676125","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

Heterogeneous Domain Remapping for Universal Detection of Generative Linguistic Steganography

IF 3.2 2区工程技术

IEEE Signal Processing Letters Pub Date : 2025-03-07 DOI: 10.1109/LSP.2025.3549015

Tong Xiao;Jingang Wang;Songbin Li

{"title":"Heterogeneous Domain Remapping for Universal Detection of Generative Linguistic Steganography","authors":"Tong Xiao;Jingang Wang;Songbin Li","doi":"10.1109/LSP.2025.3549015","DOIUrl":"https://doi.org/10.1109/LSP.2025.3549015","url":null,"abstract":"Current researchers have proposed various steganalysis methods for detecting secret information within social media texts, which can achieve relatively optimal detection performance in specific steganographic domains. However, considering the practical application of social media, we can only obtain the text to be tested without prior knowledge of the steganographic domain it belongs to. Consequently, we are unable to prepare a supervised training dataset in advance. This places higher demands on steganalysis algorithms, necessitating their ability to generalize and detect any unknown steganography domain. To this end, we propose a universal detection method for generative linguistic steganography based on heterogeneous domain remapping. The core idea is to employ a neural structure composed of pre-trained embedding layers and capsule networks to extract steganography-sensitive correlation features. Subsequently, the concept of contrastive learning is utilized to remap the sensitive features from heterogeneous steganography domains into a unified domain. This process effectively extracts domain-invariant features, thereby enabling the detection of unknown steganographic domains. Experimental results demonstrate that the proposed method outperforms existing approaches by an average of over 2% across various steganography domains.","PeriodicalId":13154,"journal":{"name":"IEEE Signal Processing Letters","volume":"32 ","pages":"1281-1285"},"PeriodicalIF":3.2,"publicationDate":"2025-03-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143716365","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