检索结果-内蒙古大学图书馆

How to Design a Cheap Music detection System Using a Simple Multilayer Perceptron With Temporal Integration

IEEE SIGNAL PROCESSING MAGAZINE 2024年第4期41卷 83-88页

作者： Rafii, Zafar Wold, Erling Boulderstone, Richard Audible Mag Los Gatos CA 95032 USA

We show how to design a cheap system for detecting when music is present in audio recordings. We make use of a small neural network consisting of a simple multilayer perceptron (MLP) along with compact features derive... 详细信息

关键词： Neural networks Multilayer perceptrons Audio recording Multiple signal classification Music Design methodology detection algorithms Spectrogram

来源：评论

学校读者我要写书评

暂无评论

Wildfire detection With Deep Learning-A Case Study for the CICLOPE Project

引用

IEEE ACCESS 2024年 12卷 82095-82110页

作者： Goncalves, Afonso M. Brandao, Tomas Ferreira, Joao C. Lovelytics Arlington VA 22201 USA ISTAR Inst Univ Lisboa ISCTE IUL P-1649026 Lisbon Portugal ISTAR INOV INESC Inovacao Inst Novas Tecnol P-1000029 Lisbon Portugal

In recent years, Portugal has seen wide variability in wildfire damage associated to high unpredictability of climatic events such as severe heatwaves and drier summers. Therefore, timely and accurate detection of forest and rural wildfires is of great importance for successful fire containment and suppression efforts, as wildfires exponentially increase their spread rate from the moment of ignition. In the field of early smoke detection, the CICLOPE project currently trailblazes in the employment of a network of Remote Acquisition Towers for wildfire prevention and observation, along with a rule-based automatic smoke detection system, covering over 2, 700, 000 hectares of wildland and rural area in continental Portugal. However, the inherent challenges of automatic smoke detection raise issues of high false alarm rates that affect the system's prediction quality and overwhelm the Management and Control Centers with numerous false alarms. The research work presented in this paper evaluates the potential improvement in wildfire smoke detection accuracy and specificity using deep learning-based architectures. It proposes a solution based on a Dual-Channel CNN that can be deployed as a secondary prediction confirmation layer to further refine the CICLOPE automatic smoke detection system. The proposed solution takes advantage of the high true alarm coverage of the current detection system by taking only the predicted alarm images and respective bounding box coordinates as inputs. The Dual-Channel network combines the widely used DenseNet architecture with a novel detail selective network with spatial and channel attention modules trained separately with image data obtained from CICLOPE, fusing the extracted features from both networks in a concatenation layer. The results demonstrate that the proposed Dual-Channel CNN outperforms both single-channel networks, achieving an accuracy of 99.7% and a low false alarm rate of 0.20% when re-examining the alarms produced by the

关键词： Feature extraction Wildfires Computer architecture Image classification Computational modeling YOLO Computer vision Convolutional neural networks Deep learning detection algorithms convolutional neural networks deep learning smoke detection wildfire detection

来源：评论

学校读者我要写书评

暂无评论

Self-Supervised Graph Transformer for Deepfake detection

引用

IEEE ACCESS 2024年 12卷 58114-58127页

作者： Khormali, Aminollah Yuan, Jiann-Shiun Univ Cent Florida Dept Elect & Comp Engn Orlando FL 32816 USA

Deepfake detection methods have shown promising results in recognizing forgeries within a given dataset, where training and testing take place on the in-distribution dataset. However, their performance deteriorates significantly when presented with unseen samples. As a result, a reliable deepfake detection system must remain impartial to forgery types, appearance, and quality for guaranteed generalizable detection performance. Despite various attempts to enhance cross-dataset generalization, the problem remains challenging, particularly when testing against common post-processing perturbations, such as video compression or blur. Hence, this study introduces a deepfake detection framework, leveraging a self-supervised pre-training model that delivers exceptional generalization ability, withstanding common corruptions and enabling feature explainability. The framework comprises three key components: a feature extractor based on vision Transformer architecture that is pre-trained via self-supervised contrastive learning methodology, a graph convolution network coupled with a Transformer discriminator, and a graph Transformer relevancy map that provides a better understanding of manipulated regions and further explains the model's decision. To assess the effectiveness of the proposed framework, several challenging experiments are conducted, including in-data distribution performance, cross-dataset & cross-manipulation generalization, and robustness against common post-production perturbations. The results achieved demonstrate the remarkable effectiveness of the proposed deepfake detection framework, surpassing the current state-of-the-art approaches.

关键词： Deepfakes Transformers Feature extraction Self-supervised learning Visualization Task analysis Rendering (computer graphics) detection algorithms Convolutional neural networks Deepfake detection graph convolution networks self-supervised contrastive learning vision transformer

来源：评论

学校读者我要写书评

暂无评论

A Novel Early detection and Prevention of Coronary Heart Disease Framework Using Hybrid Deep Learning Model and Neural Fuzzy Inference System

引用

IEEE ACCESS 2024年 12卷 26683-26695页

作者： Ramesh, B. Lakshmanna, Kuruva Vellore Inst Technol Sch Comp Sci Engn & Informat Syst Vellore 632014 Tamil Nadu India

Diabetes is the "mother of all diseases" as it affects multiple organs of body of an individual in some way. Its timely detection and management are critically important. Otherwise, the long run, it can cause several complications in a diabetic. Heart disease is one of the major complications of *** work proposed an Optimal Scrutiny Boosted Graph Convolutional LSTM (O-SBGC-LSTM), SBGC-LSTM enhanced by Eurygaster Optimization Algorithm (EOA) to tune hyperparameters for early prevention and detection of diabetes disease. This work proposed an Optimal Scrutiny Boosted Graph Convolutional LSTM (O-SBGC-LSTM), SBGC-LSTM enhanced by Eurygaster Optimization Algorithm (EOA) to tune hyperparameters for early prevention and detection of diabetes disease. This method not only captures discriminative features in spatial configuration and temporal dynamics but also explore the co-occurrence relationship between spatial and temporal domains. This method also presents a temporal hierarchical architecture to increase temporal receptive fields of top SBGC-LSTM layer, which boosts the ability to learn high-level semantic representation and significantly reduces computation cost. The performance of O-SBGC-LSTM was found overall to be satisfactory, reaching >98% accuracy in most studies. In comparison with classic machine learning approaches, proposed hybrid DL was found to achieve better performance in almost all studies that reported such comparison outcomes. Furthermore, prevention is better than cure. Additionally, employed fuzzy based inference techniques to enhance the prevention procedure using suggestion table.

关键词： Diabetes Diseases Heart Fuzzy logic Predictive models Prediction algorithms Training detection algorithms Coronary arteriosclerosis Convolutional neural networks Long short term memory Fuzzy systems Inference algorithms detection prevention coronary heart disease optimal scrutiny boosted graph convolutional LSTM deep learning fuzzy inference systems

来源：评论

学校读者我要写书评

暂无评论

Fully Synthetic Pedestrian Anomaly Behavior Dataset Generation in Metaverse for Enhancing Autonomous Driving Object detection

引用

IEEE ACCESS 2024年 12卷 166630-166642页

作者： Aung, Nang Htet Htet Sangwongngam, Paramin Jintamethasawat, Rungroj Wuttisittikulkij, Lunchakorn Chulalongkorn Univ Fac Engn Wireless Commun Ecosyst Res Unit Bangkok 10330 Thailand Natl Sci & Technol Dev Agcy NSTDA Natl Elect & Comp Technol Ctr NECTEC Khlong Luang 12120 Pathum Thani Thailand

Pedestrian detection is fundamental in the realm of autonomous driving, relying on comprehensive datasets for accurate deep-learning methods. Detecting diverse pedestrian behaviors, including rare and near-accident cases, is necessary, however, there were limitations in the development of such specific datasets. Furthermore, involving humans in data collection is risky and raises privacy concerns since recording and storing data necessitates explicit consent from all participants. To tackle this challenge, we propose MetaPed, a synthetic pedestrian dataset generated from the Metaverse, where the avatars can represent a wide range of pedestrian behaviors without exposing real individuals to potential harm or privacy violations. Our Metaverse includes both programmed pedestrian behaviors and controllable avatar movements and interactions, ensuring a comprehensive and realistic representation of pedestrian actions. To validate and ensure the generalization capabilities of our synthetic dataset's performance in real-world scenarios, we perform both cross-dataset and intra-dataset evaluations on comprehensive real-world datasets, including KITTI, Citypersons, and INRIA Person. Our experiments demonstrate that the network trained on our synthetic dataset exhibits robust generalization capabilities for unseen real-world situations and our dataset's inclusion significantly enhanced the Average Precision (AP) and Average Recall (AR) of the pedestrian detection model. For cross-dataset validation, we generally observe improvements in both AP and AR with models trained on all datasets. The inclusion of our dataset also improves the intra-dataset validation of the INRIA Person dataset, where the AP and AR values increase from 0.958 to 0.988 and AR from 0.697 to 0.725, respectively. For Citypersons and KITTI datasets, the AP values increase from 0.630 to 0.639 and 0.745 to 0.752, respectively. The model training on our dataset achieved the highest AP and AR values across all tes

关键词： Pedestrians Synthetic data Training Metaverse Object detection Data privacy Meteorology Data models Autonomous vehicles Visualization detection algorithms Autonomous vehicles (AVs) deep learning pedestrian detection pedestrian anomaly behaviors synthetic dataset metaverse dataset validation

来源：评论

学校读者我要写书评

暂无评论

ZDDR: A Zero-Shot Defender for Adversarial Samples detection and Restoration

引用

IEEE ACCESS 2024年 12卷 39081-39094页

作者： Chen, Musheng He, Guowei Wu, Junhua Guangdong Vocat Coll Inst Elect Informat Guangzhou 510800 Peoples R China

Natural language processing (NLP) models find extensive applications but face vulnerabilities against adversarial inputs. Traditional defenses lean heavily on supervised detection techniques, which makes them vulnerable to issues arising from training data quality, inherent biases, noise, or adversarial inputs. This study observed common compromises in sentence fluency during aggression. On this basis, the Zero Sample Defender (ZDDR) is introduced for adversarial sample detection and recovery without relying on prior knowledge. ZDDR combines the log probability calculated by the model and the syntactic normative score of a large language model (LLM) to detect adversarial examples. Furthermore, using strategic prompts, ZDDR guides LLM in rephrasing adversarial content, maintaining clarity, structure, and meaning, thereby restoring the sentence from the attack. Benchmarking reveals a 9% improvement in area under receiver operating characteristic curve (AUROC) for adversarial detection over existing techniques. Post-restoration, model classification efficacy surges by 45% compared to the offensive inputs, setting new performance standards against other restoration techniques.

关键词： Training Robustness Computational modeling Perturbation methods Semantics Natural language processing Data models Adversarial machine learning detection algorithms Adversarial defense large language model natural language processing model security prompt engineering

来源：评论

学校读者我要写书评

暂无评论

A Survey on IoT Ground Sensing Systems for Early Wildfire detection: Technologies, Challenges, and Opportunities

引用

IEEE ACCESS 2024年 12卷 172785-172819页

作者： Chan, Chiu Chun Alvi, Sheeraz A. Zhou, Xiangyun Durrani, Salman Wilson, Nicholas Yebra, Marta Australian Natl Univ Sch Engn Canberra ACT 2601 Australia Australian Natl Univ Bushfire Res Ctr Excellence Canberra ACT 2601 Australia Australian Natl Univ Fenner Sch Environm & Soc Canberra ACT 2601 Australia

The threat posed by wildfires or bushfires has become a severe global issue due to the increase in human activities in forested areas and the impact of climate change. Consequently, there is a surge in the development of automatic wildfire detection methods. Approaches based on long-distance imagery from satellites or watchtowers encounter limitations, such as restricted visibility, which results in delayed response times. To address and overcome these challenges, research interest has grown in the implementation of ground-based Internet of Things (IoT) sensing systems for early wildfire detection. However, research on energy consumption, detection latency, and detection accuracy of IoT sensing systems, as well as the performance of various anomaly detection algorithms when evaluated using these metrics, is lacking. Therefore, in this article, we present an overview of current IoT ground sensing systems for early wildfire detection. Camera and environmental sensing technologies suitable for early wildfire detection are discussed, as well as vision-based detection algorithms and detection algorithms for environmental sensing. Challenges related to the development and implementation of IoT ground sensing systems for early wildfire detection together with suggestions for creating a robust detection system to combat the growing threat of wildfires worldwide are discussed.

关键词： Climate change Fires Forests Internet of Things Sensors detection algorithms Ground support Bushfire early fire detection forest fire Internet of Things (IoT) sensing wildfire

来源：评论

学校读者我要写书评

暂无评论

Fully Automated Deep Learning-Based Renal Mass detection on Multi-Parametric MRI

引用

IEEE ACCESS 2024年 12卷 112714-112728页

作者： Gaikar, Rohini Azad, Azar Schieda, Nicola Ukwatta, Eranga Univ Guelph Sch Engn Guelph ON N1G 2W1 Canada AI VALI Inc Toronto ON M4N 3N1 Canada Univ Ottawa Dept Radiol Ottawa ON K1H 8L6 Canada

Due to superior soft tissue contrast afforded by magnetic resonance imaging (MRI), there is great potential for multi-parametric MRI (mpMRI) for the detection and eventual classification of renal masses (RMs). In this study, we investigated fully automated deep learning methods for RMs detection using T2-Weighted (T2W) spin-echo and two contrast-enhanced T1-Weighted gradient-echo-corticomedullary (T1W-CM), nephrographic-phase (T1W-NG), T1-Weighted In-phase (T1W-IP) and opposed-phase (T1W-OP) images. The dataset contained mpMRI images of 108 kidney cancer patients with an average size of renal mass of 24 +/- 7.8 cm. In the first stage, kidneys were segmented using a 2D attention U-Net model, which was reported in a previous study. In the second stage, we tested five different state-of-the-art methods for RMs detections on mpMRI sequences. The model predictions were compared to manual annotations using precision, recall, specificity, and Dice Similarity Coefficient (DSC). The best-performing deep learning models were U-Net, U-Net++, and attention U-Net on the T2W, T1W-CM, and T1W-NG sequences respectively. Of the 5 mpMRI sequences, we also demonstrated that the T1W-CM is the most suitable for RMs detection. This automated detection of RMs in mpMRI sequences may be useful for the subsequent characterization of RMs in a fully automated artificial intelligence-based pipeline.

关键词： Kidney Image segmentation Magnetic resonance imaging Cancer Training Manuals detection algorithms Deep learning Biomedical imaging Multi-parametric MRI kidney cancer computer-aided detection of cancer deep learning

来源：评论

学校读者我要写书评

暂无评论

TrellisBsP: Trellis-Based Belief-Selective Propagation for MIMO detection

引用

IEEE WIRELESS COMMUNICATIONS LETTERS 2025年第4期14卷 939-943页

作者： Tan, Zeqiong Zhou, Wenyue Huang, Yongming Zhang, Chuan You, Xiaohu Southeast Univ Natl Mobile Commun Res Lab LEADS Nanjing 211189 Peoples R China Southeast Univ Frontiers Sci Ctr Mobile Informat Commun & Secur Nanjing 211189 Peoples R China Purple Mt Labs Nanjing 211100 Peoples R China

Compared to its linear counterparts, the belief propagation (BP) MIMO detector achieves better performance in challenging MIMO scenarios. For practical applications, the recently emerged belief-selective propagation (BsP) detector mitigates the error floor issue and reduces the complexity. However, it still suffers from intractable complexity for high-order MIMO scenarios, and the serial message-updating mechanism limits its hardware feasibility. In this letter, a trellis-based BsP (TrellisBsP) detection algorithm is proposed. Two unexplored approaches: trellis representation and pathfinding strategy are utilized to enable parallel message-updating while achieving complexity reduction. Benchmarking with the BsP detection, the proposed TrellisBsP saves up to 72.76% complexity in an 8 x 4, 256-QAM MIMO system without error performance penalty. In addition, the proposed TrellisBsP has great potential in hardware implementation due to its parallel message-updating mechanism.

关键词： Symbols Hardware Detectors Vectors Reliability Computational complexity detection algorithms Decoding Buildings Uplink Multiple-input multiple-output (MIMO) detection belief-selective propagation (BsP) trellis representation pathfinding strategy

来源：评论

学校读者我要写书评

暂无评论

YOLOv8n-ASF-DH: An Enhanced Safety Helmet detection Method

引用

IEEE ACCESS 2024年 12卷 126313-126328页

作者： Lin, Bingyan Fujian Polytech Informat Technol Fuzhou 350003 Peoples R China

Wearing a safety helmet is an essential aspect of safety production management. To address the issue of low detection accuracy of existing safety helmet detection algorithms for small targets and complex scenes, we propose the YOLOv8n-ASF-DH model (DH stands for Dynamic Head), which is based on YOLOv8n and integrates the Attentional Scale Sequence Fusion (ASF) structure and the Dynamic Head (DyHead). In the backbone layer, a Triplet Attention mechanism is incorporated to enhance the model's focus on features of small targets. In the neck layer, the ASF structure efficiently fuses different level output features extracted by the backbone network, enhancing the overall feature representation capability of the model. In the detection head, DyHead adjusts the relationships between different feature layers, enhancing the model's performance on different scales and complex scenes. Additionally, adopting the Focal-EIoU loss function balances the contributions of high-quality and low-quality samples in loss calculation. Experimental results demonstrate that the improved model enhances detection performance in complex scenes and small target scenarios. Compared to the YOLOv8n model, the YOLOv8n-ASF-DH model achieved an improvement of 3.066% in accuracy and 3.883% in recall. Additionally, the model exhibited an increase of 2.584% in mAP 0.5 and 6.131% in mAP 0.5:0.95. Compared with other mainstream object detection algorithms, the improved model significantly enhances the performance of safety helmet target detection tasks.

关键词： Feature extraction Accuracy detection algorithms YOLO Support vector machines Safety devices Head-mounted displays Helmet wearing detection YOLOv8 attentional scale sequence fusion dynamic head triplet attention mechanism

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：