Pub Date : 2024-11-20DOI: 10.1109/LSP.2024.3503359
Kai Wang;Mingliang Zhou;Qing Lin;Guanglin Niu;Xiaowei Zhang
Point cloud completion 3D object detectors effectively tackle the challenge of incomplete shapes in sparse point clouds by generating pseudo points to improve detection performance. However, the absence of guidance provided by the heatmap information and the geometric shape information renders the precise recovery of object shapes an arduous task. To this end, we propose a Geometry-guided Point Generation for 3D Object Detection, named GgPG. Specifically, we first design a 3D heatmap auxiliary supervision subnetwork to enhance the quality of object proposals by capturing the actual size and position of the object within the 3D heatmap representation. Moreover, we introduce a density-aware point generation module that employs Kernel Density Estimation (KDE) to embed the point density into the grid point's feature representation, thereby enabling the completion of more precise object shapes. Our GgPG achieves progressive performance in both Waymo and KITTI benchmarks, notably GgPG outperforms PGRCNN by +1.02 $%$ , +1.18 $%$ , and +0.56 $%$ on the vehicle, pedestrian, and cyclist under LEVEL $_$ 2 mAPH classes on Waymo Open Dataset, respectively.
{"title":"Geometry-Guided Point Generation for 3D Object Detection","authors":"Kai Wang;Mingliang Zhou;Qing Lin;Guanglin Niu;Xiaowei Zhang","doi":"10.1109/LSP.2024.3503359","DOIUrl":"https://doi.org/10.1109/LSP.2024.3503359","url":null,"abstract":"Point cloud completion 3D object detectors effectively tackle the challenge of incomplete shapes in sparse point clouds by generating pseudo points to improve detection performance. However, the absence of guidance provided by the heatmap information and the geometric shape information renders the precise recovery of object shapes an arduous task. To this end, we propose a Geometry-guided Point Generation for 3D Object Detection, named GgPG. Specifically, we first design a 3D heatmap auxiliary supervision subnetwork to enhance the quality of object proposals by capturing the actual size and position of the object within the 3D heatmap representation. Moreover, we introduce a density-aware point generation module that employs Kernel Density Estimation (KDE) to embed the point density into the grid point's feature representation, thereby enabling the completion of more precise object shapes. Our GgPG achieves progressive performance in both Waymo and KITTI benchmarks, notably GgPG outperforms PGRCNN by +1.02\u0000<inline-formula><tex-math>$%$</tex-math></inline-formula>\u0000, +1.18\u0000<inline-formula><tex-math>$%$</tex-math></inline-formula>\u0000, and +0.56\u0000<inline-formula><tex-math>$%$</tex-math></inline-formula>\u0000 on the vehicle, pedestrian, and cyclist under LEVEL\u0000<inline-formula><tex-math>$_$</tex-math></inline-formula>\u0000 2 mAPH classes on Waymo Open Dataset, respectively.","PeriodicalId":13154,"journal":{"name":"IEEE Signal Processing Letters","volume":"32 ","pages":"136-140"},"PeriodicalIF":3.2,"publicationDate":"2024-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142844509","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}
Pub Date : 2024-11-20DOI: 10.1109/LSP.2024.3503365
Mingye Ge;Yatao Liu;Mingjie Shao;Haixia Zhang
The future machine-type communication in internet-of-things (IoT) systems involves a massive number of devices sporadically communicating with a base station (BS) equipped with multiple antennas. Detecting active devices and estimating their associated channels are crucial but challenging due to the large number of potential devices and the small fraction of active devices. Existing studies assume high-resolution analog-to-digital converters (ADCs) at the BS, while there is a growing interest in implementing low-resolution ADCs, particularly one-bit ADCs, in massive multiple-input multiple-output (MIMO) systems. This paper focuses on the joint one-bit active device detection and channel estimation problem. We consider the maximum-likelihood approach and propose a novel expectation maximization (EM) algorithm with acceleration. On the theoretical aspect, we provide the convergent computational complexity analysis for the accelerated EM algorithm. The proposed method, evaluated through numerical simulations, outperforms benchmark algorithms in terms of both estimation accuracy and computational complexity.
{"title":"Maximum-Likelihood Active Device Detection and Channel Estimation in One-Bit MIMO System","authors":"Mingye Ge;Yatao Liu;Mingjie Shao;Haixia Zhang","doi":"10.1109/LSP.2024.3503365","DOIUrl":"https://doi.org/10.1109/LSP.2024.3503365","url":null,"abstract":"The future machine-type communication in internet-of-things (IoT) systems involves a massive number of devices sporadically communicating with a base station (BS) equipped with multiple antennas. Detecting active devices and estimating their associated channels are crucial but challenging due to the large number of potential devices and the small fraction of active devices. Existing studies assume high-resolution analog-to-digital converters (ADCs) at the BS, while there is a growing interest in implementing low-resolution ADCs, particularly one-bit ADCs, in massive multiple-input multiple-output (MIMO) systems. This paper focuses on the joint one-bit active device detection and channel estimation problem. We consider the maximum-likelihood approach and propose a novel expectation maximization (EM) algorithm with acceleration. On the theoretical aspect, we provide the convergent computational complexity analysis for the accelerated EM algorithm. The proposed method, evaluated through numerical simulations, outperforms benchmark algorithms in terms of both estimation accuracy and computational complexity.","PeriodicalId":13154,"journal":{"name":"IEEE Signal Processing Letters","volume":"32 ","pages":"141-145"},"PeriodicalIF":3.2,"publicationDate":"2024-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142844295","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}
Given partial measurements of a time-varying graph signal, we propose an algorithm to simultaneously estimate both the underlying graph topology and the missing measurements. The proposed algorithm operates by training an interpretable neural network, designed from the unrolling framework. The proposed technique can be used as a graph learning and/or a graph signal reconstruction algorithm. This work builds on prior work in graph learning by tailoring the learned graph to the signal reconstruction task; and also enhances prior work in graph signal reconstruction by allowing the underlying graph to be unknown.
{"title":"Inpainting-Driven Graph Learning via Explainable Neural Networks","authors":"Subbareddy Batreddy;Pushkal Mishra;Yaswanth Kakarla;Aditya Siripuram","doi":"10.1109/LSP.2024.3501273","DOIUrl":"https://doi.org/10.1109/LSP.2024.3501273","url":null,"abstract":"Given partial measurements of a time-varying graph signal, we propose an algorithm to simultaneously estimate both the underlying graph topology and the missing measurements. The proposed algorithm operates by training an interpretable neural network, designed from the unrolling framework. The proposed technique can be used as a graph learning and/or a graph signal reconstruction algorithm. This work builds on prior work in graph learning by tailoring the learned graph to the signal reconstruction task; and also enhances prior work in graph signal reconstruction by allowing the underlying graph to be unknown.","PeriodicalId":13154,"journal":{"name":"IEEE Signal Processing Letters","volume":"32 ","pages":"111-115"},"PeriodicalIF":3.2,"publicationDate":"2024-11-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142825959","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}
Pub Date : 2024-11-18DOI: 10.1109/LSP.2024.3501286
Xiaorong Xu;Shuo Yang;Zhaoting Liu;Jun Wu;Jianrong Bao
This letter investigates the design and optimization of secure beamforming in an intelligent reflecting surface (IRS)-assisted multiuser cognitive radio simultaneous wireless information and power transfer (CR-SWIPT) system. The proposed method leverages IRS to address cognitive energy harvesting nodes as potential eavesdroppers (Eve). The objective is to maximize the achievable secrecy rate while satisfying multiple constraints, such as transmit power control, energy harvesting, phase shifts, and maximum tolerable interference power. To solve this highly non-convex optimization problem, we propose a hybrid algorithm that combines successive convex approximation (SCA) and inexact block coordinate descent (IBCD). By decomposing the problem into three sub-problems, local optimal beamforming matrices and phase-shift matrices are obtained using the SCA method and complex circle manifold (CCM) method, respectively. Simulation results show that the secrecy rate at the SWIPT information decoding node in the IRS-assisted multiuser CR-SWIPT system improves significantly, with an approximate 40% enhancement compared to the random phase shift scheme with maximum transmit power. The effectiveness of the proposed algorithm is further validated through secrecy rate performance evaluations under different system configurations.
{"title":"SCA and IBCD Hybrid Algorithm Based Secure Beamforming Optimization for IRS-Assisted Multiuser CR-SWIPT System","authors":"Xiaorong Xu;Shuo Yang;Zhaoting Liu;Jun Wu;Jianrong Bao","doi":"10.1109/LSP.2024.3501286","DOIUrl":"https://doi.org/10.1109/LSP.2024.3501286","url":null,"abstract":"This letter investigates the design and optimization of secure beamforming in an intelligent reflecting surface (IRS)-assisted multiuser cognitive radio simultaneous wireless information and power transfer (CR-SWIPT) system. The proposed method leverages IRS to address cognitive energy harvesting nodes as potential eavesdroppers (Eve). The objective is to maximize the achievable secrecy rate while satisfying multiple constraints, such as transmit power control, energy harvesting, phase shifts, and maximum tolerable interference power. To solve this highly non-convex optimization problem, we propose a hybrid algorithm that combines successive convex approximation (SCA) and inexact block coordinate descent (IBCD). By decomposing the problem into three sub-problems, local optimal beamforming matrices and phase-shift matrices are obtained using the SCA method and complex circle manifold (CCM) method, respectively. Simulation results show that the secrecy rate at the SWIPT information decoding node in the IRS-assisted multiuser CR-SWIPT system improves significantly, with an approximate 40% enhancement compared to the random phase shift scheme with maximum transmit power. The effectiveness of the proposed algorithm is further validated through secrecy rate performance evaluations under different system configurations.","PeriodicalId":13154,"journal":{"name":"IEEE Signal Processing Letters","volume":"32 ","pages":"336-340"},"PeriodicalIF":3.2,"publicationDate":"2024-11-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142890273","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}
Pub Date : 2024-11-18DOI: 10.1109/LSP.2024.3501272
Marcelo A. Colominas;Sylvain Meignen
This paper aims at building a new instantaneous frequency (IF) estimator of the modes making up non-stationary multi-component signals, using reassignment operators used in Fourier-based synchrosqueezing transforms (FSSTs) and linear chirp points detection. Reassignment operators provide with different IF estimates, depending on the assumption made on the local polynomial order of the phase of the studied mode. In practice, it is difficult to estimate locally which order fits the best, and to choose too high an order, typically larger than two, when not necessary results in both an inaccurate estimation and an increased sensitivity to noise. To circumvent this, we propose to localize linear chirp points in modes to find out where second order phase approximation is sufficient for the estimation, and then build a new IF estimate based on a weighted spline approximation based on these points. Numerical results show the improvement brought by the proposed approach in noisy situations over classical IF estimators used in FSSTs.
{"title":"Instantaneous Frequency Estimation Based on Reassignment Operators and Linear Chirp Points Detection","authors":"Marcelo A. Colominas;Sylvain Meignen","doi":"10.1109/LSP.2024.3501272","DOIUrl":"https://doi.org/10.1109/LSP.2024.3501272","url":null,"abstract":"This paper aims at building a new instantaneous frequency (IF) estimator of the modes making up non-stationary multi-component signals, using reassignment operators used in Fourier-based synchrosqueezing transforms (FSSTs) and linear chirp points detection. Reassignment operators provide with different IF estimates, depending on the assumption made on the local polynomial order of the phase of the studied mode. In practice, it is difficult to estimate locally which order fits the best, and to choose too high an order, typically larger than two, when not necessary results in both an inaccurate estimation and an increased sensitivity to noise. To circumvent this, we propose to localize linear chirp points in modes to find out where second order phase approximation is sufficient for the estimation, and then build a new IF estimate based on a weighted spline approximation based on these points. Numerical results show the improvement brought by the proposed approach in noisy situations over classical IF estimators used in FSSTs.","PeriodicalId":13154,"journal":{"name":"IEEE Signal Processing Letters","volume":"32 ","pages":"106-110"},"PeriodicalIF":3.2,"publicationDate":"2024-11-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142825845","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}
Pub Date : 2024-11-18DOI: 10.1109/LSP.2024.3501274
Hangjie Yi;Jingsheng Qian;Yuhang Ming;Wanzeng Kong
The Steady State Visual Evoked Potential (SSVEP) paradigm has been widely employed in various Brain-Computer Interface (BCI) systems. However, recent studies indicate that SSVEP is vulnerable to adversarial attacks, resulting in manipulated results and drastic degradation in recognition performance, which pose inconveniences and even risks to users. Noticing the fact that the adversarial attack on SSVEP is done by adding subtle waveform perturbations into random EEG channels, we propose Independent Components Time-Frequency Purification with Channel Consensus (ICTFP-CC) as a defensive strategy. In particular, we first detect and remove suspicious perturbations with independent component analysis from the time and frequency domain, and then reconstruct the purified EEG signals. Additionally, we introduce a voting mechanism to achieve channel consensus and enhance overall robustness. We conducted experiments on two public datasets and three SSVEP recognition algorithms. The results demonstrate that our method can significantly improve the classification accuracy and information transfer rate of attacked SSVEP signals by a maximum of 46.79 (%) and 62.87 (bits/min).
{"title":"Independent Components Time-Frequency Purification With Channel Consensus Against Adversarial Attack in SSVEP-Based BCIs","authors":"Hangjie Yi;Jingsheng Qian;Yuhang Ming;Wanzeng Kong","doi":"10.1109/LSP.2024.3501274","DOIUrl":"https://doi.org/10.1109/LSP.2024.3501274","url":null,"abstract":"The Steady State Visual Evoked Potential (SSVEP) paradigm has been widely employed in various Brain-Computer Interface (BCI) systems. However, recent studies indicate that SSVEP is vulnerable to adversarial attacks, resulting in manipulated results and drastic degradation in recognition performance, which pose inconveniences and even risks to users. Noticing the fact that the adversarial attack on SSVEP is done by adding subtle waveform perturbations into random EEG channels, we propose Independent Components Time-Frequency Purification with Channel Consensus (ICTFP-CC) as a defensive strategy. In particular, we first detect and remove suspicious perturbations with independent component analysis from the time and frequency domain, and then reconstruct the purified EEG signals. Additionally, we introduce a voting mechanism to achieve channel consensus and enhance overall robustness. We conducted experiments on two public datasets and three SSVEP recognition algorithms. The results demonstrate that our method can significantly improve the classification accuracy and information transfer rate of attacked SSVEP signals by a maximum of 46.79 (%) and 62.87 (bits/min).","PeriodicalId":13154,"journal":{"name":"IEEE Signal Processing Letters","volume":"32 ","pages":"116-120"},"PeriodicalIF":3.2,"publicationDate":"2024-11-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142825908","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}
Pub Date : 2024-11-18DOI: 10.1109/LSP.2024.3501957
Xinyu Xiong;Wenxue Li;Jie Ma;Duojun Huang;Siying Li
Existing semi-supervised polyp segmentation methods assume that unlabeled images are positive, containing lesions to be annotated, while neglecting negative samples that are widely available in practice. This letter reveals that harvesting lesion-free negative samples can effectively boost polyp segmentation performance. Directly extending the labeled set with negative samples is sub-optimal since it introduces potential class imbalance. To overcome this challenge, we first introduce a data augmentation strategy named TypeMix. By fusing unlabeled samples with negative samples, the network can better benefit from diverse features provided by negatives while alleviating the potential side effects. Furthermore, it is observed that the number of negative samples significantly exceeds that of lesion samples. To reduce redundancy and improve training efficiency, we propose a dynamic informativeness-aware sampling strategy, prioritizing the active selection of high-valuable negative samples. Extensive experiments on public datasets demonstrate that our simple but effective strategies are enough to consistently outperform other state-of-the-art methods, offering new possibilities for future work from a data collection perspective.
{"title":"Free Meal: Boosting Semi-Supervised Polyp Segmentation by Harvesting Negative Samples","authors":"Xinyu Xiong;Wenxue Li;Jie Ma;Duojun Huang;Siying Li","doi":"10.1109/LSP.2024.3501957","DOIUrl":"https://doi.org/10.1109/LSP.2024.3501957","url":null,"abstract":"Existing semi-supervised polyp segmentation methods assume that unlabeled images are positive, containing lesions to be annotated, while neglecting negative samples that are widely available in practice. This letter reveals that harvesting lesion-free negative samples can effectively boost polyp segmentation performance. Directly extending the labeled set with negative samples is sub-optimal since it introduces potential class imbalance. To overcome this challenge, we first introduce a data augmentation strategy named TypeMix. By fusing unlabeled samples with negative samples, the network can better benefit from diverse features provided by negatives while alleviating the potential side effects. Furthermore, it is observed that the number of negative samples significantly exceeds that of lesion samples. To reduce redundancy and improve training efficiency, we propose a dynamic informativeness-aware sampling strategy, prioritizing the active selection of high-valuable negative samples. Extensive experiments on public datasets demonstrate that our simple but effective strategies are enough to consistently outperform other state-of-the-art methods, offering new possibilities for future work from a data collection perspective.","PeriodicalId":13154,"journal":{"name":"IEEE Signal Processing Letters","volume":"32 ","pages":"131-135"},"PeriodicalIF":3.2,"publicationDate":"2024-11-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142844571","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}
Pub Date : 2024-11-18DOI: 10.1109/LSP.2024.3501285
Shuangbing Wen;Qishan Zhang;Tao Hu;Jun Li
Artificial intelligence technology has been developing rapidly, and speech synthesis models have become increasingly mature, capable of generating highly realistic synthetic audio used to disseminate misinformation, which poses a serious security risk problem. Digital watermarking technology can effectively protect digital content. Deep learning is currently achieving significant research success in digital watermarking. However, the current robustness against audio manipulation remains understudied. Based on this, we propose a robust audio watermarking method based on deep learning against manipulation attacks. Specifically, the embedding of watermarking information is performed in the encoder and the extraction of watermarking information is performed in the decoder; In addition, various audio attacks are simulated during iterative training, a sampling noise layer is used to increase robustness, and a discriminator is used to distinguish between encoded audio and original audio to improve the invisibility of the watermark. We comprehensively evaluate the performance of our model against various manipulation attacks. Experimental results demonstrate that the framework effectively embeds and extracts watermarked signals, exhibiting strong robustness.
{"title":"Robust Audio Watermarking Against Manipulation Attacks Based on Deep Learning","authors":"Shuangbing Wen;Qishan Zhang;Tao Hu;Jun Li","doi":"10.1109/LSP.2024.3501285","DOIUrl":"https://doi.org/10.1109/LSP.2024.3501285","url":null,"abstract":"Artificial intelligence technology has been developing rapidly, and speech synthesis models have become increasingly mature, capable of generating highly realistic synthetic audio used to disseminate misinformation, which poses a serious security risk problem. Digital watermarking technology can effectively protect digital content. Deep learning is currently achieving significant research success in digital watermarking. However, the current robustness against audio manipulation remains understudied. Based on this, we propose a robust audio watermarking method based on deep learning against manipulation attacks. Specifically, the embedding of watermarking information is performed in the encoder and the extraction of watermarking information is performed in the decoder; In addition, various audio attacks are simulated during iterative training, a sampling noise layer is used to increase robustness, and a discriminator is used to distinguish between encoded audio and original audio to improve the invisibility of the watermark. We comprehensively evaluate the performance of our model against various manipulation attacks. Experimental results demonstrate that the framework effectively embeds and extracts watermarked signals, exhibiting strong robustness.","PeriodicalId":13154,"journal":{"name":"IEEE Signal Processing Letters","volume":"32 ","pages":"126-130"},"PeriodicalIF":3.2,"publicationDate":"2024-11-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142844570","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}
Early and accurate polyp segmentation is crucial for the diagnosis and treatment of colorectal cancer. However, polyp segmentation faces many challenges: different polyp sizes, complex shapes, and ambiguous intestinal wall boundaries. To solve these problems, we propose a novel polyp segmentation network named DeformSegNet. Specifically, we first introduce a polyp perception module (PPM), which combines the dynamic multi-kernel spatial selection network (DMS-Net) and a transformer encoder to effectively locate polyps of different sizes. Next, we design a deformation-aware separable module (DSM), which consists of deformable attention that adaptively adjusts the sampling position, enabling the network to adapt to complex and diverse polyp boundaries. Finally, a cross-attention aggregation module (CAAM) effectively retains low-level features, further enhancing the boundary features and suppressing false positives. DeformSegNet achieves competitive segmentation accuracy on five polyp datasets, demonstrating excellent learning and generalization capabilities.
{"title":"Boundary Refinement Network for Polyp Segmentation With Deformable Attention","authors":"Zijian Li;Zhiyong Yang;Wangsheng Wu;Zihang Guo;Dongdong Zhu","doi":"10.1109/LSP.2024.3501283","DOIUrl":"https://doi.org/10.1109/LSP.2024.3501283","url":null,"abstract":"Early and accurate polyp segmentation is crucial for the diagnosis and treatment of colorectal cancer. However, polyp segmentation faces many challenges: different polyp sizes, complex shapes, and ambiguous intestinal wall boundaries. To solve these problems, we propose a novel polyp segmentation network named DeformSegNet. Specifically, we first introduce a polyp perception module (PPM), which combines the dynamic multi-kernel spatial selection network (DMS-Net) and a transformer encoder to effectively locate polyps of different sizes. Next, we design a deformation-aware separable module (DSM), which consists of deformable attention that adaptively adjusts the sampling position, enabling the network to adapt to complex and diverse polyp boundaries. Finally, a cross-attention aggregation module (CAAM) effectively retains low-level features, further enhancing the boundary features and suppressing false positives. DeformSegNet achieves competitive segmentation accuracy on five polyp datasets, demonstrating excellent learning and generalization capabilities.","PeriodicalId":13154,"journal":{"name":"IEEE Signal Processing Letters","volume":"32 ","pages":"121-125"},"PeriodicalIF":3.2,"publicationDate":"2024-11-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142825952","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}
We consider the problem of recovering off-the-grid spikes from linear measurements in the context of Single Molecule Localization Microscopy (SMLM). State of the art model-based methods such as Over-Parametrized Continuous Orthogonal Matching Pursuit (OP-COMP) with Projected Gradient Descent (PGD) have been shown to successfully recover those signals. The computational cost of these methods scales linearly with the number of measurements. When this number of measurements is large with respect to the dimensionality of the signal, we propose to reduce it with a so-called sketching operator. Based on recent results on compressive sensing in the space of measures, we approximate the ideal sketching operator (benefiting from theoretical recovery guarantees), in the context of SMLM. This sketching method coupled to OP-COMP with PGD shows significant improvements in calculation time in realistic synthetic microscopy experiments.
{"title":"Sketched Over-Parametrized Projected Gradient Descent for Sparse Spike Estimation","authors":"Pierre-Jean Bénard;Yann Traonmilin;Jean-François Aujol","doi":"10.1109/LSP.2024.3488496","DOIUrl":"https://doi.org/10.1109/LSP.2024.3488496","url":null,"abstract":"We consider the problem of recovering off-the-grid spikes from linear measurements in the context of Single Molecule Localization Microscopy (SMLM). State of the art model-based methods such as Over-Parametrized Continuous Orthogonal Matching Pursuit (OP-COMP) with Projected Gradient Descent (PGD) have been shown to successfully recover those signals. The computational cost of these methods scales linearly with the number of measurements. When this number of measurements is large with respect to the dimensionality of the signal, we propose to reduce it with a so-called sketching operator. Based on recent results on compressive sensing in the space of measures, we approximate the ideal sketching operator (benefiting from theoretical recovery guarantees), in the context of SMLM. This sketching method coupled to OP-COMP with PGD shows significant improvements in calculation time in realistic synthetic microscopy experiments.","PeriodicalId":13154,"journal":{"name":"IEEE Signal Processing Letters","volume":"32 ","pages":"16-20"},"PeriodicalIF":3.2,"publicationDate":"2024-11-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142825909","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}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：