检索结果-内蒙古大学图书馆

IEEE International Conference on Field-Programmable technology (FPT)

作者： Ning Zhang Shuo Ni Tingting Qiao Wenchao Liu He Chen Beijing Key Laboratory of Embedded Real-time Information Processing Technology Beijing Institute of Technology Beijing China

Directly completing remote sensing scene classification (RSSC) on space platforms can minimize latency and relieve data downlink burdens. The all adder neural network $( \mathrm{A}^{2}$NN) is a novel network for on-board RSSC with lower resource overhead than convolutional neural networks (CNNs). However, most of the existing FPGA-based accelerators are designed for CNNs and are not applicable to deploy $\mathrm{A}^{2}$NNs. In this paper, we propose an extremely pipelined FPGA-based accelerator of $\mathrm{A}^{2}$NNs and implement a VGGNet-11 backbone for on-board RSSC. In the proposed FPGA-based accelerator, an extremely pipelined processing engine (PE) suitable for accelerating the adder layer is designed. Each adder layer in the $\mathrm{A}^{2}$NN is mapped to a dedicated extremely pipelined PE, which achieves low-latency calculations. Besides, the entire parameters of the network are stored in block RAMs, and the intermediate data are cached on the FPGA chip, thereby eliminating external memory accesses and reducing power consumption. To evaluate the performance of the proposed extremely pipelined FPGA-based accelerator of $\mathrm{A}^{2}$NN, we implemented an $\mathrm{A}^{2}$NN-based RSCC model with the VGGNet-11 backbone on the Xilinx Virtex7 XC7VX690T FPGA by the proposed accelerator. The experimental results show that the proposed FPGA-based accelerator of $\mathrm{A}^{2}$NNs can achieve a throughput of 3.04 tera operations per second (TOPs) at 200 MHz while consuming 8.27 W.

关键词：

来源：评论

学校读者我要写书评

暂无评论

SAR TOMOGRAPHY IN URBAN AREAS: AN APPROACH BASED ON SPARSE SPECTRUM FITTING

SAR TOMOGRAPHY IN URBAN AREAS: AN APPROACH BASED ON SPARSE S...

引用

IET International Radar Conference 2023, IRC 2023

作者： Zhao, Jian Wang, Zhen Ding, Zegang Sun, Tao Chen, Zhizhou Chen, Haolong Zeng, Tao Radar Research Lab School of Information and Electronics Beijing Institute of Technology Beijing100081 China Beijing Key Laboratory of Embedded Real-Time Information Processing Technology Beijing Institute of Technology Beijing100081 China Chongqing Innovation Center Beijing Institute of Technology Chongqing401331 China

ISBN: (纸本)9781839539954

Tomography Synthetic Aperture Radar (TomoSAR) technology, as an extension of Interferometric SAR (InSAR), solves the inherent layover problem and provides three-dimensional (3D) imaging ability. Now, it is a promising technique in urban management and critical situation detection. However, in practical applications, the sparse and non-uniform distribution characteristics of the repeat-pass baselines make TomoSAR face the problem of low resolution and high sidelobes, which affect the 3D imaging effect. To address the above problems, a sparse spectrum fitting (SpSF) based TomoSAR processing approach is proposed in this paper. Compared with the traditional spatial spectrum estimation method, this method can achieve higher resolution and lower sidelobes, which is a significant advantage. The proposed method is also validated using computer simulation and the TerraSAR-X real data. © The Institution of Engineering & technology 2023.

关键词： Radar imaging

来源：评论

学校读者我要写书评

暂无评论

Multi-Grained Global-Local Semantic Feature Fusion for Few Shot Remote Sensing Scene Classification

Multi-Grained Global-Local Semantic Feature Fusion for Few S...

引用

IEEE International Symposium on Geoscience and Remote Sensing (IGARSS)

作者： Yuqing Liu Tong Zhang Yin Zhuang Guanqun Wang He Chen Beijing Key Laboratory of Embedded Real-Time Information Processing Technology Beijing Institute of Technology Beijing China

Few-shot remote sensing scene classification aims to classify unseen scenes by using only a few labeled samples. Hence, how to set up a more effective feature description according to a few labeled samples, becomes an important issue. In this paper, in view of more complicated remote sensing scenes containing several hierarchical and coupled spatial relations (e.g., internal and external spatial contexts), which severely hinder the feature extraction under few-shot learning scenarios, a multi-grained global-local semantic feature fusion (MGGL-SFF) method is proposed for few-shot remote sensing scene classification, which can better combine the global discriminative spatial semantic features with local transferable fragment features to set a powerful prototype representation up for few shot learning. Finally, experiments are carried out on defined few-shot remote sensing scene classification benchmark, and results proved the proposed MGGL-SFF can achieve a new state-of-the-art performance.

关键词：

来源：评论

学校读者我要写书评

暂无评论

An FPGA-GPU Heterogeneous System and Implementation for On-Board Remote Sensing Data processing

An FPGA-GPU Heterogeneous System and Implementation for On-B...

引用

IEEE International Conference on Field-Programmable technology (FPT)

作者： Tingting Qiao Yu Xie He Chen Yizhuang Xie Beijing Key Laboratory of Embedded Real-time Information Processing Technology Beijing Institute of Technology Beijing China

In resource-constrained space environments, energy-efficient and high-performance processors like Field- Programmable Gate Arrays (FPGAs) and Digital Signal Processors (DSPs) are widely utilized for on-board data processing. However, the variety of tasks in space missions, including remote sensing data pre-processing and processing, reveals the limitations of single processors. To address this, we suggest a heterogeneous system that combines FPGAs with embedded Graphics processing Unit (GPU) TX2 for efficient remote sensing data pre-processing and processing. Additionally, to overcome the challenges in data transfer between the FPGA and TX2, we utilized High-Level Synthesis (HLS) to design a dedicated IP for efficient data transfer between the FPGA and TX2 using PCIe protocol. This method enhances data transfer speeds and enables TX2 to manage address mapping in the Double Data Rate (DDR) memory. Our system successfully completes remote sensing data processing by integrating these elements.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Partial Triple Modular Redundancy Method for Fault-Tolerant Circuit based on HITS Algorithm

Partial Triple Modular Redundancy Method for Fault-Tolerant ...

引用

IEEE International Symposium on Defect and Fault Tolerance in VLSI Systems

作者： Yu Xie Wen-Yue Yu Ning Zhang He Chen Yi-Zhuang Xie Beijing Key Laboratory of Embedded Real-Time Information Processing Technology Beijing Institute of Technology Beijing China

Circuits related to nuclear facilities and on-orbit circuits typically require fault-tolerant reinforcement to cope with harsh radiation environments. While the traditional fully Triple Modular Redundancy (TMR) method can effectively reinforce circuits, it suffers from the problems of large circuit area and high power consumption. To solve this problem, this paper proposes a Partial TMR method based on the HITS algorithm (Hyperlink-Induced Topic Search), which can effectively perform circuit fault tolerance. This method sorts the importance of triggers in the circuit based on the HITS algorithm and uses the Fault Interval time (MTBF) of the Bi-Logic Cone model as the judgment criterion to determine the number of triggers that need to be redundantly designed under partial TMR. Some commonly used modules of an on-orbit processor were tested to verify the effectiveness of the proposed redundancy method in this paper. The experimental results show that the resource overhead can be reduced by at least 20% while achieving an MTBF of 99.9% under TMR.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A tracking filter algorithm for migratory birds based on group situation constraints

A tracking filter algorithm for migratory birds based on gro...

引用

Signal, information and Data processing (ICSIDP), IEEE International Conference on

作者： Biao Li Rui Wang Jiong Cai The Radar Research Lab School of Information and Electronics Beijing Institute of Technology Beijing Key Laboratory of Embedded Real-Time Information Processing Technology Beijing China

ISBN: (数字)9798331515669

ISBN: (纸本)9798331515676

Aerial animal migration is closely related to human life, and radar is one of the most effective means of observing it. However, scanning radar has a low data rate, a short integration time, and lacks prior knowledge of target speed, while migratory animals exhibit high mobility. These factors make traditional tracking filter algorithms ineffective at predicting their movement. To address this issue, this paper proposes a tracking filter algorithm based on group dynamics constraints, leveraging the characteristic behavior of migratory animals flying in groups. First, the dynamic state of the group is calculated using the group as the research unit. Then, tracking filtering is performed under the constraint of group dynamics. Simulation and field data have verified the effectiveness of the proposed algorithm.

关键词： Target tracking Heuristic algorithms Dynamics Radar Filtering algorithms Radar tracking Prediction algorithms Birds information filters Radar signal processing

来源：评论

学校读者我要写书评

暂无评论

SAR-DARTS: A CONVOLUTIONAL NEURAL ARCHITECTURE SEARCH FOR SAR SCENE CLASSIFICATION

SAR-DARTS: A CONVOLUTIONAL NEURAL ARCHITECTURE SEARCH FOR SA...

引用

IET International Radar Conference 2023, IRC 2023

作者： Zhou, Tian Xu, Yinsheng Li, Qihai Qi, Baogui Chen, He Chen, Liang Beijing Key Laboratory of Embedded Real-time Information Processing Technology Beijing Institute of Technology Beijing100081 China Yangtze Delta Region Academy of Beijing Institute of Technology Zhejiang Jiaxing314019 China Shanghai Institute of Satellite Engineering Shanghai200240 China

ISBN: (纸本)9781839539954

In recent years, the collaborative service architecture of cloud computing and edge computing to reduce the demand for computing, storage and other capabilities of artificial intelligence methods has developed rapidly, but their application in synthetic aperture radar (SAR) image scene classification is still immature. In this paper, we propose a neural architecture search and pruning model for SAR scene classification. By introducing network morphisms and path regularization method, we can reduce the unfair competition between cells on the basis of expanding the search space. The model approach is applied to the FUSAR-Ship and OpenSARShip datasets for performance evaluation. The proposed method can achieve higher accuracy with fewer parameters. © The Institution of Engineering & technology 2023.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Simultaneous segmentation and classification of colon cancer polyp images using a dual branch multi-task learning network

引用

Mathematical Biosciences and Engineering 2024年第2期21卷 2024-2049页

作者： Li, Chenqian Liu, Jun Tang, Jinshan School of Computer Science and Technology Wuhan University of Science and Technology Wuhan430065 China Hubei Province Key Laboratory of Intelligent Information Processing and Real-Time Industrial System Wuhan430065 China Department of Health Administration and Policy College of Public Health George Mason University FairfaxVA22030 United States

Accurate classification and segmentation of polyps are two important tasks in the diagnosis and treatment of colorectal cancers. Existing models perform segmentation and classification separately and do not fully make use of the correlation between the two tasks. Furthermore, polyps exhibit random regions and varying shapes and sizes, and they often share similar boundaries and backgrounds. However, existing models fail to consider these factors and thus are not robust because of their inherent limitations. To address these issues, we developed a multi-task network that performs both segmentation and classification simultaneously and can cope with the aforementioned factors effectively. Our proposed network possesses a dual-branch structure, comprising a transformer branch and a convolutional neural network (CNN) branch. This approach enhances local details within the global representation, improving both local feature awareness and global contextual understanding, thus contributing to the improved preservation of polyp-related information. Additionally, we have designed a feature interaction module (FIM) aimed at bridging the semantic gap between the two branches and facilitating the integration of diverse semantic information from both branches. This integration enables the full capture of global context information and local details related to polyps. To prevent the loss of edge detail information crucial for polyp identification, we have introduced a reverse attention boundary enhancement (RABE) module to gradually enhance edge structures and detailed information within polyp regions. Finally, we conducted extensive experiments on five publicly available datasets to evaluate the performance of our method in both polyp segmentation and classification tasks. The experimental results confirm that our proposed method outperforms other state-of-the-art methods. ©2024 the Author(s), licensee AIMS Press.

关键词： Classification (of information)

来源：评论

学校读者我要写书评

暂无评论

A MODEL COMPRESSION METHOD BASED ON MULTI-TEACHER DISTILLATION FOR SAR SCENE CLASSIFICATION

A MODEL COMPRESSION METHOD BASED ON MULTI-TEACHER DISTILLATI...

引用

IET International Radar Conference 2023, IRC 2023

作者： Li, Qihai Zhang, Jiawei Zhou, Tian Qi, Baogui Chen, He Chen, Liang Beijing Key Laboratory of Embedded Real-time Information Processing Technology Beijing Institute of Technology Beijing100081 China Yangtze Delta Region Academy of Beijing Institute of Technology Zhejiang Jiaxing314019 China Shanghai Institute of Satellite Engineering Shanghai200240 China

ISBN: (纸本)9781839539954

In recent years, with the advancement of deep learning, convolutional neural networks tailored for processing image content have gained increasing attention within the remote sensing imagery domain. However, deep convolutional neural networks that exhibit exceptional performance often encompass a substantial number of parameters, thereby necessitating heightened computational resources and time for network training and prediction. Given the constraints posed by resource limitations in the realm of remote sensing, deploying such high-performance networks onto these devices becomes a challenging endeavor. To address this set of challenges, this paper introduces a methodology grounded in target differential probability for multi-teacher knowledge distillation. This approach effectively reduces the model's complexity while preserving the network models' commendable performance. By conducting a statistical analysis of the model's output probability distribution and computing the differential probabilities between teacher and student networks, we regulate the output proportions of teachers within the multi-teacher model. As a result, a high-performing student network is cultivated. In aggregate, experimental validation on two distinct datasets substantiates the efficacy of this approach in facilitating efficient knowledge transfer, consequently yielding heightened performance outcomes. © The Institution of Engineering & technology 2023.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

FIMD: Fusion-Inspired Modality Distillation for Enhanced MRI Segmentation in Incomplete Multi-Modal Scenarios 24

FIMD: Fusion-Inspired Modality Distillation for Enhanced MRI...

引用

7th International Conference on Image and Graphics processing, ICIGP 2024

作者： Yang, Rui Wang, Xiao Xu, Xin Hubei Province Key Laboratory of Intelligent Information Processing and Real-Time Industrial System Wuhan University of Science and Technology China School of Computer Science and Technology Wuhan University of Science and Technology China Hubei Provincial Key Laboratory of Multimedia and Network Communication Engineering Wuhan China

ISBN: (纸本)9798400716720

Magnetic Resonance Imaging (MRI) is an invaluable tool for brain tumor segmentation. However, in clinical practice, certain modalities might be unavailable, leading to potential performance degradation in prediction tasks. According to current implementations, different modalities are treated as independent entities during the feature extraction phase, ignoring their inherent complementary nature. In this paper, inspired by knowledge distillation techniques, we introduce the Fusion-Inspired Modality Distillation (FIMD) framework. FIMD leverages the output from a fused multi-modal model as a teacher to guide the training of individual modalities. Each modality benefits from the collective knowledge of all modalities, enhancing its performance. Our FIMD method, devoid of specific architectural constraints, seamlessly integrates into existing multi-modal brain tumor segmentation frameworks. Notably, extensive experiments on BraTS2020, BraTS2018, and BraTS2015 datasets indicate that FIMD can enhance the performance of current state-of-the-art (SOTA) algorithms on Dice by an average of 0.55-1.60%. Especially in the case of missing modalities, the effectiveness of the FIMD framework in enhancing the robustness and accuracy of brain tumor segmentation is demonstrated. © 2024 ACM.

关键词： Semantic Segmentation

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：