检索结果-内蒙古大学图书馆

5th International Conference on deep learning, Artificial Intelligence and Robotics, ICDLAIR 2023

作者： Tiwari, Kanishka Bhatia, Anmolpreet Singh Garg, Nibhrit Arora, Ishaan Saini, Poonam Department of Computer Science and Engineering Punjab Engineering College Chandigarh India

ISBN: (纸本)9783031609343

The utilization of web-based applications has greatly increased in recent years. This, in turn, has raised security concerns and data breaches leading to losses of millions of dollars. A large part of these attacks is carried out through modifications in HTTP requests and are often too complex to be recognised through simple rule-based checks. Thus, it is vital to utilize machine learning approaches in order to protect the applications against such attack vectors. Additionally, given the pace at which web servers need to operate, these machine learning methods need to be extremely fast at processing the requests. The paper presents a comparison of two recent deep learning models to recognize the web application attacks namely, a convolution neural network (CNN) and a Transformer-based model. We compare the prediction accuracies of both the models as well as their prediction time to identify the best model suited for being deployed in a real-life scenario. Based on implementation and results, it is observed that for our dataset, CNN tends to perform better in terms of both, accuracy and prediction time. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

MultiPanoWise: holistic deep architecture for multi-task dense prediction from a single panoramic image

MultiPanoWise: holistic deep architecture for multi-task den...

引用

IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

作者： Shah, Uzair Tukur, Muhammad Alzubaidi, Mahmood Pintore, Giovanni Gobbetti, Enrico Househ, Mowafa Schneider, Jens Agus, Marco Hamad Bin Khalifa Univ ICT Div Coll Sci & Engn Doha Qatar CRS4 Visual & Data Intens Comp Pula Italy Natl Res Ctr HPC Big Data & Quantum Comp Pula Italy

ISBN: (纸本)9798350365474

We present a novel holistic deep-learning approach for multi-task learning from a single indoor panoramic image. Our framework, named MultiPanoWise, extends vision transformers to jointly infer multiple pixel-wise signals, such as depth, normals, and semantic segmentation, as well as signals from intrinsic decomposition, such as reflectance and shading. Our solution leverages a specific architecture combining a transformer-based encoder-decoder with multiple heads, by introducing, in particular, a novel context adjustment approach, to enforce knowledge distillation between the various signals. Moreover, at training time we introduce a hybrid loss scalarization method based on an augmented Chebychev/hypervolume scheme. We illustrate the capabilities of the proposed architecture on publicdomain synthetic and real-world datasets. We demonstrate performance improvements with respect to the most recent methods specifically designed for single tasks, like, for example, individual depth estimation or semantic segmentation. To our knowledge, this is the first architecture capable of achieving state-of-the-art performance on the joint extraction of heterogeneous signals from single indoor omnidirectional images.

关键词： dense estimation indoor environments multi-task learning panoramic images

来源：评论

学校读者我要写书评

暂无评论

FPGA Based Military Vehicle Classification from Drone-Based Video Using deep learning 9

FPGA Based Military Vehicle Classification from Drone-Based ...

引用

9th International Conference on Frontiers of Signal processing, ICFSP 2024

作者： Vasavi, S. Sree Sowmya, D. Aishwarya, Ch Fuentes, Wendy Flores Velagapudi Ramakrishna Siddhartha Engineering College Computer Science & Enigneering Vijayawada India Engineering Institute Autonomous University of Baja California Mexico

ISBN: (纸本)9798350353235

Remotely controlled aerial vehicles such as drones are used for military applications such as surveillance, intelligence and target acquisition. real-time object detection and classification is one of the most recurrent tasks for drones. The proposed system is used to classify the military vehicles from the drone-based videos. Object segmentation and classification of military vehicles (military tanks and APCs) is done using the Detectron2 and is based on Mask-RCNN benchmark. A custom dataset consisting of 300 images of Military Tanks and Armored Personnel Carriers (APCs) has been created and annotated for the purpose of the project. Furthermore, the Mask RCNN (Resnet50+FPN) model is trained with the custom dataset to classify the vehicles. The trained model is integrated in to Field Programmable Gate Arrays (FPGA), to carry out onboard classification in a real-time manner. Performance measures are used to evaluate the proposed model and overall accuracy is 93% with reduced false positive rate. © 2024 IEEE.

关键词： Drones

来源：评论

学校读者我要写书评

暂无评论

Deere learning approach for hyperspectral image demosaicking, spectral correction and high-resolution RGB reconstruction

引用

COMPUTER METHODS IN BIOMECHANICS AND BIOMEDICAL ENGINEERING-IMAGING AND VISUALIZATION 2022年第4期10卷 409-417页

作者： Li, Peichao Ebner, Michael Noonan, Philip Horgan, Conor Bahl, Anisha Ourselin, Sebastien Shapey, Jonathan Vercauteren, Tom Kings Coll London Sch Biomed Engn & Imaging Sci London England Hypervis Surg Ltd London England Kings Coll Hosp NHS Fdn Trust Dept Neurosurg London England

Hyperspectral imaging is one of the most promising techniques for intraoperative tissue characterisation. Snapshot mosaic cameras, which can capture hyperspectral data in a single exposure, have the potential to make a real-time hyperspectral imaging system for surgical decision-making possible. However, optimal exploitation of the captured data requires solving an ill-posed demosaicking problem and applying additional spectral corrections. In this work, we propose a supervised learning-based image demosaicking algorithm for snapshot hyperspectral images. Due to the lack of publicly available medical images acquired with snapshot mosaic cameras, a synthetic image generation approach is proposed to simulate snapshot images from existing medical image datasets captured by high-resolution, but slow, hyperspectral imaging devices. image reconstruction is achieved using convolutional neural networks for hyperspectral image super-resolution, followed by spectral correction using a sensor-specific calibration matrix. The results are evaluated both quantitatively and qualitatively, showing clear improvements in image quality compared to a baseline demosaicking method using linear interpolation. Moreover, the fast processing time of 45 ms of our algorithm to obtain super-resolved RGB or oxygenation saturation maps per image for a state-of-the-art snapshot mosaic camera demonstrates the potential for its seamless integration into real-time surgical hyperspectral imaging applications.

关键词： Hyperspectral imaging demosaicking deep learning surgical imaging

来源：评论

学校读者我要写书评

暂无评论

Accurate and Efficient Loop Closure Detection With deep Binary image Descriptor and Augmented Point Cloud Registration

Accurate and Efficient Loop Closure Detection With Deep Bina...

引用

2024 International Conference on Intelligent Robots and Systems

作者： Wang, Jialiang Gao, Zhi Lin, Zhipeng Zhou, Zhiyu Wang, Xiaonan Cheng, Jianhua Zhang, Hao Liu, Xinyi Chen, Ben M. Wuhan Univ Sch Remote Sensing & Informat Engn Wuhan 430072 Peoples R China Chinese Univ Hong Kong Dept Mech & Automat Engn Hong Kong 999077 Peoples R China Wuhan Univ Hubei Luojia Lab Wuhan 430079 Peoples R China ZG Technol Co Ltd Wuhan 430000 Peoples R China

ISBN: (纸本)9798350377712;9798350377705

Loop Closure Detection (LCD) is an essential component of Simultaneous Localization and Mapping (SLAM), helping to correct drift errors, facilitate map merging, or both by identifying previously observed scenes. Despite its importance, traditional LCD algorithms based on single sensor such as camera or LiDAR exhibit degraded performance in challenging scenarios due to their inherent limitations. To address this issue, we propose a novel LCD method based on camera-LiDAR fusion, exploiting the rich textural information from cameras and the accurate geometric data from LiDAR to ensure robustness and speed in challenging environments. Specifically, we first employ deep hashing learning to encode deep image features into binary image descriptors for extremely fast loop candidate (LC) retrieval. Then, LiDAR points are augmented with image color for accurate geometric verification. Finally, we incorporate a spatial-temporal consistency check that mandates an LC to have consistently matched neighbors to be accepted as true. Our method is extensively verified and compared with the state-of-the-art methods on various datasets encompassing both indoor and outdoor environments. Experimental results demonstrate that our method obtains the best performance, increasing the maximum recall rate at 100% precision by a significant margin of 20% while operating in real-time at an average speed of 30 fps.

关键词： Binary images

来源：评论

学校读者我要写书评

暂无评论

An effective deep CNN-LSTM based intrusion detection system for network security

An effective deep CNN-LSTM based intrusion detection system ...

引用

2024 International Conference on Control, Automation and Diagnosis, ICCAD 2024

作者： Jablaoui, Rahma Liouane, Noureddine University of Monastir Laboratory of Automatic Signal and Image Processing Monastir Tunisia

ISBN: (纸本)9798350361025

deep learning (DL) models have become prevalent and consistently exhibit outstanding performance across diverse domains, particularly in information security. As a subset of machine learning (ML), DL has proven adept at handling complex, unstructured, and large-scale data, making it well-suited to tackle the evolving challenges of cybersecurity. By leveraging DL techniques, we can develop robust intrusion detection systems (IDS) to counter a wide range of network attacks, serving as a crucial defense against hackers and cybercriminals. The integration of DL into IDS holds immense potential for significantly enhancing network security and mitigating cyber risks in various real-world scenarios. This work proposes an IDS architecture that integrates Convolutional Neural Network (CNN) for spatial feature extraction and Long Short-Term Memory (LSTM) for temporal feature extraction and sequential analysis to predict and classify malicious cyberattacks in IoT network traffic. Our experiments with the CSE-CIC-IDS2018 dataset, the latest comprehensive network traffic dataset, demonstrate the superiority of our approach in terms of accuracy, precision, recall, and F1 score. Our model achieves an unprecedented accuracy of over 99% in both binary and multiclass classification, surpassing existing efforts in the literature. © 2024 IEEE.

关键词： Long short-term memory

来源：评论

学校读者我要写书评

暂无评论

Hybrid path planning and following of a quadrotor UAV based on deep reinforcement learning 36

Hybrid path planning and following of a quadrotor UAV based ...

引用

36th Chinese Control and Decision Conference (CCDC)

作者： Zhong, Liangsheng Zhao, Jiasheng Luo, Haining Hou, Zhiwei Sun Yat Sen Univ Sch Syst Sci & Engn Guangzhou Peoples R China

ISBN: (纸本)9798350387780;9798350387797

deep reinforcement learning(DRL)-based path planning algorithms have gained significant attention in recent years due to their end-to-end processing and robustness. Nevertheless, they face challenges in scenarios with long distances and dense obstacles because they only consider local environmental information. This paper presents a hybrid path planning and following approach that combines path planning based on soft actorcritic(SAC) with path following. Initially, a path is generated using the sampling-based method Adaptively Informed Trees (AIT*), and the path subsequently serves as tracking points for the planner. This approach guides the agent to move faster and more effectively toward the goal, and continuously updates the tracking point in real time. We conducted an experiment to evaluate the training process and performance of this hybrid path planning approach based on DRL while comparing it to the original DRL-based approach. The experimental results illustrate the superiority of the presented approach in both the training process and task performance.

关键词： Path Planning Path Follow deep Reinforcement learning UAV Soft Actor-Critic

来源：评论

学校读者我要写书评

暂无评论

image Based Detection of Cotton Plant Diseases Using Discrete Wavelet Transform Features 3

Image Based Detection of Cotton Plant Diseases Using Discret...

引用

3rd IEEE International Conference on Intelligent Techniques in Control, Optimization and Signal processing, INCOS 2024

作者： Mohan, Shalini Muthulakshmi, Lakshmanan Devaraj, D. Kalasalingam Academy of Research and Education Department of Biotechnology Srivilliputtur India Kalasalingam Academy of Research and Education Department of Electrical and Electronics Engineering Srivillliputur India

ISBN: (纸本)9798350361186

Cotton is one of the most vital cash crops cultivated around the globe and its yield directly influences the economy and the livelihood of huge number of people associated with it. The major loss of cotton occurs due to damage incurred by microbial pathogens that infect the cotton crop. The course of infection and their effect on the growth and yield of the plant varies depending upon the type of disease. Appropriate treatment for the specific type of disease could prevent the spread of infection and also reduce the use of pesticide that is usually done during the empirical treatment techniques. Identification of the type of cotton plant disease is critical and needs to be done at a faster pace. image processingbased machine learning approach is found to be a better option for this application. This article deals with development of the machine learning model capable of identifying the different types of cotton plant diseases using image processing. Foliar images of diseased cotton plants are collected from public databases and preprocessing operations like resizing, filtering and contrast enhancement are performed followed by k-means clustering for segmentation of diseased part of the leaf. image decomposition is performed using discrete wavelet transform with Daubechies wavelet at five levels and the mean, standard deviation, and entropy of the coefficients are used for building the machine learning model. Classification with Artificial Neural Network trained with backpropagation algorithm has a classification accuracy of 97% indicating the veracity of the model. Hence, the developed model could be used as for developing a user-friendly interface that could be tapped for real-time applications. © 2024 IEEE.

关键词： Discrete wavelet transforms

来源：评论

学校读者我要写书评

暂无评论

Quantized spiral-phase-modulation based deep learning for real-time defocusing distance prediction

引用

OPTICS EXPRESS 2022年第15期30卷 26931-26940页

作者： Zhang, Zezheng Chan, Ryan K. Y. Wong, Kenneth K. Y. Univ Hong Kong Dept Elect & Elect Engn Pokfulam Rd Hong Kong Peoples R China Adv Biomed Instrumentat Ctr Shatin Hong Kong Sci Pk Hong Kong Peoples R China

Whole slide imaging (WSI) has become an essential tool in pathological diagnosis, owing to its convenience on remote and collaborative review. However, how to bring the sample at the optimal position in the axial direction and image without defocusing artefacts is still a challenge, as traditional methods are either not universal or time-consuming. Until recently, deep learning has been shown to be effective in the autofocusing task in predicting defocusing distance. Here, we apply quantized spiral phase modulation on the Fourier domain of the captured images before feeding them into a light-weight neural network. It can significantly reduce the average predicting error to be lower than any previous work on an open dataset. Also, the high predicting speed strongly supports it can be applied on an edge device for real-time tasks with limited computational source and memory footprint. (C) 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement

关键词： High numerical aperture optics image processing Neural networks Phase modulation Spiral phase Stochastic gradient descent

来源：评论

学校读者我要写书评

暂无评论

A Quantum Spatial Graph Convolutional Neural Network Model on Quantum Circuits

引用

IEEE TRANSACTIONS ON NEURAL NETWORKS AND learning SYSTEMS 2024年第6期PP卷 9837-9851页

作者： Zheng, Jin Gao, Qing Ogorzalek, Maciej Lu, Jinhu Deng, Yue Beihang Univ Sch Automat Sci & Elect Engn Beijing 100191 Peoples R China Zhongguancun Lab Beijing 100094 Peoples R China Beihang Univ Hangzhou Innovat Inst Hangzhou 310000 Peoples R China Jagiellonian Univ Dept Informat Technol PL-30348 Krakow Poland Beihang Univ Sch Automat Sci & Elect Engn Beijing Peoples R China Beihang Univ Sch Astronaut Beijing 102206 Peoples R China

This article proposes a quantum spatial graph convolutional neural network (QSGCN) model that is implementable on quantum circuits, providing a novel avenue to processing non-Euclidean type data based on the state-of-the-art parameterized quantum circuit (PQC) computing platforms. Four basic blocks are constructed to formulate the whole QSGCN model, including the quantum encoding, the quantum graph convolutional layer, the quantum graph pooling layer, and the network optimization. In particular, the trainability of the QSGCN model is analyzed through discussions on the barren plateau phenomenon. Simulation results from various types of graph data are presented to demonstrate the learning, generalization, and robustness capabilities of the proposed quantum neural network (QNN) model.

关键词： Graph convolutional neural networks (GCNs) quantum neural networks (QNNs) quantum spatial graph convolutional neural network (QSGCN)

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：