检索结果-内蒙古大学图书馆

Attention mechanism combined with residual recurrent neural network for sound event detection and localization

EURASIP JOURNAL ON AUDIO SPEECH AND MUSIC PROCESSING 2022年第1期2022卷 29-29页

作者： Lan, Chaofeng Zhang, Lei Zhang, Yuanyuan Fu, Lirong Sun, Chao Han, Yulan Zhang, Meng Harbin Univ Sci & Technol Dept Sch Measurement & Commun Engn Harbin 150080 Peoples R China Beidahuang Ind Grp Gen Hosp Harbin 150088 Peoples R China Hainan Univ Mech & Elect Engn Coll Haikou 570228 Peoples R China Guangzhou Univ Sch Elect & Commun Engn Guangzhou 510006 Peoples R China

In the task of sound event detection and localization (SEDL) in a complex environment, the acoustic signals of different events usually have nonlinear superposition, so the detection and localization effect is not good. Given this, this paper is based on the Residual-spatially and channel Squeeze-Excitation (Res-scSE) model. Combined with Multiple-scale Convolutional Recurrent Neural Network (M-CRNN), the Res-scSE-CRNN model is proposed. Firstly, to solve the problem of insufficient extraction of time-frequency feature in single-size convolution kernel, multi-scale feature fusion is carried out by using the feature hierarchy of the convolutional neural network to improve the accuracy of detection. Secondly, aiming at the problem of overlapping audio event localization accuracy is not high, with Res-scSE to replace common convolution module and add residual structure to strengthen the feature extraction, and combining with an attention mechanism to enhance neural network channels and spatial relationships, to improve the network to extract the characteristics of directivity, achieve the goal of the overlapped audio localization. In this paper, experiments are carried out in the open dataset DCASE2019, and evaluation indicators are used to analyze the effectiveness of the proposed model and baseline model in the detection and localization of audio events. The results show that compared with the M-CRNN model, the detection error rate of Res-scSE-CRNN model is reduced 4%, the F1-Score is increased 3.4%, the localization error is reduced by 22.8 & DEG;, and the frame recall rate is increased 3%.

关键词： Sound event detection and localization Convolutional cyclic neural network Multi-scale feature fusion Space channel squeeze excitation module

来源：评论

学校读者我要写书评

暂无评论

The equipment detection and localization of large-scale construction jobsite by far-field construction surveillance video based on improving YOLOv3 and grey wolf optimizer improving extreme learning machine

引用

CONSTRUCTION AND BUILDING MATERIALS 2021年 291卷 123268-123268页

作者： Zeng, Tuocheng Wang, Jiajun Cui, Bo Wang, Xiaoling Wang, Dong Zhang, Yunuo Tianjin Univ State Key Lab Hydraul Engn Simulat & Safety Tianjin 300072 Peoples R China

The equipment detection and localization on large-scale construction jobsite by far-field construction surveillance video is significant to monitor jobsite safety and construction process. These videos usually with a larger view containing extensive construction equipment and richer environment information could help the supervisor know more about real-time status of the large-scale construction jobsite. However, due to numerous multi-scale objects, especially for small scale objects in the video, it is difficult to recognize each equipment intuitively, which is not conducive to ensuring jobsite security and guiding construction process. This study presents a real-time multi-scale equipment detection and localization model including the improving YOLOv3 for detection and the grey wolf optimizer improving extreme learning machine (GWO-ELM) for localization. The effectiveness and efficiency of the proposed model are demonstrated via a real hydraulic-engineering megaproject. The experimental results show that the improving YOLOv3 has improved 4.8% mean average precision (mAP) over the multi-scale objects detection, especially, the significant improvements of 17.2% mAP and 13.3% recall at small scale objects detection while maintaining real-time. The improving YOLOv3 is proved robust under various environmental conditions on construction monitoring. Meanwhile, the mean absolute localization percentage errors of different scale objects are less than 1%, which indicates the accuracy and effectiveness of the model to assist supervisor in inspecting construction safety and process. (C) 2021 Elsevier Ltd. All rights reserved.

关键词： Far-field construction surveillance video and image detection and localization Real-time Multi-scale Large-scale construction jobsite Deep learning

来源：评论

学校读者我要写书评

暂无评论

Generalized Dechirp-Keystone Transform for Radar High-Speed Maneuvering Target detection and localization

引用

REMOTE SENSING 2021年第17期13卷 3367-3367页

作者： Zheng, Jibin Zhu, Kangle Niu, Zhiyong Liu, Hongwei Liu, Qing Huo Xidian Univ Natl Lab Radar Signal Proc Xian 710071 Peoples R China Zhengzhou Univ Light Ind Coll Comp & Commun Engn Zhengzhou 450001 Peoples R China Duke Univ Dept Elect Engn Durham NC 27708 USA

The multivariate range function of the high-speed maneuvering target induces modulations on both the envelop and phase, i.e., the range cell migration (RCM) and Doppler frequency migration (DFM) which degrade the long-time coherent integration used for detection and localization. To solve this problem, many long-time coherent integration methods have been proposed. Based on mechanisms of typical methods, this paper names two signal processing modes, i.e., processing unification (PU) mode and processing separation (PS) mode, and presents their general forms. Thereafter, based on the principle of the PS mode, a novel long-time coherent integration method, known as the generalized dechirp-keystone transform (GDKT), is proposed for radar high-speed maneuvering target detection and localization. The computational cost, energy integration, peak-to-sidelobe level (PSL), resolution, and anti-noise performance of the GDKT are analyzed and compared with those of the maximum likelihood estimation (MLE) method and keystone transform-dechirp (KTD) method. With mathematical analyses and numerical simulations, we validate two main superiorities of the GDKT, including (1) the statistically optimal anti-noise performance, and (2) the low computational cost. The real radar data is also used to validate the GDKT. It is worthwhile noting that, based on closed analytical formulae of the MLE method, KTD method, and GDKT, several doubts in radar high-speed maneuvering target detection and localization are mathematically interpreted, such as the blind speed sidelobe (BSSL) and the relationship between the PU and PS modes.

关键词： detection and localization long-time coherent integration range cell migration Doppler frequency migration blind speed sidelobe

来源：评论

学校读者我要写书评

暂无评论

Laser-Induced Fluorescence Bronchoscopy for detection and localization of Early Lung Cancer

引用

Chinese Journal of Lasers 2001年第4期 70-76页

作者： XIE Shusen 1,2 LI Buhong 1,2 LI Hui 1 ZHENG Wei 1,3 LU Zukang 2 1) Biomedical Optical Lab, Institute of Laser, Fujian Teachers University, Fuzhou 35007, China 2) National Research Center of Optical Instrument Engineering, Zh

Experimenatal results on the development of a Laser Induced Fluorescence Bronchoscopy(LIFB) for the detection and localization of early lung cancer are reported in this paper. The system utilizes fluorescence of photosensitizer drug to provide real time video imaging for the examined lung tissue. Color filters are used to differentiate signal from background and a computer image processing technique is also applied to subtract the background. Moreover, a pseudocolor contrast enhancement method was developed to enhance the fluorescence image displayed on the vidio monitor. Suspicious areas are identified by pseudocolor image to guide biopsy, and several clinical trials show that sensitivity and contrast capability of the system should permit the detection and localization of early lung cancer.

关键词： laser induced fluorescence early lung cancer detection and localization computer image processing

来源：评论

学校读者我要写书评

暂无评论

INTEGRATION OF COMPLEMENTARY detection localization SYSTEMS - AN EXAMPLE

引用

IEEE TRANSACTIONS ON INFORMATION THEORY 1994年第3期40卷 808-819页

作者： KADOTA, TT AT&T BELL LABS MURRAY HILLNJ 07974

The integration (or fusion) of two or more detection-localization systems often results in performance improvement, especially if they are complementary to each other. On the other hand, the cost of integration in additional signal processing and overhead may be considerable. As a figure of merit for this tradeoff, we propose the dB-gain-to-cost ratio and illustrate its evaluation by using a simple example. The example consists of a pair of detection-localization systems which have complementary features. One has higher resolution that the other, but is more susceptible to random disturbances. We model the first as a system which detects a signal in multiplicative and additive random disturbances and additive white noise. Thus, one main result is the derivation of the maximum likelihood detector under these disturbances. In the second system, we assume the random disturbances to be negligible. Hence, the maximum likelihood detector is the matched filter. Input data to both systems are two-dimensional arrays of real numbers corresponding to the same physical area under observation where the second data array has a coarser grid than the first, thus yielding poorer resolution. The detection probabilities, in terms of which the dB-gains are computed, are evaluated by Monte Carlo simulation.

关键词： detection THEORY detection and localization detection IN NON-GAUSSIAN NOISE SURVEILLANCE SYSTEMS SYSTEM INTEGRATION

来源：评论

学校读者我要写书评

暂无评论

Marine Mammal Monitoring on Navy Ranges (M3R): A Toolset for Automated detection, localization, and Monitoring of Marine Mammals in Open Ocean Environments

引用

MARINE TECHNOLOGY SOCIETY JOURNAL 2014年第1期48卷 5-20页

作者： Jarvis, Susan M. Morrissey, Ronald P. Moretti, David J. DiMarzio, Nancy A. Shaffer, Jessica A. Naval Undersea Warfare Ctr Div Newport RI 02841 USA

Navy sonar has been associated with a number of marine mammal stranding events worldwide. As a result determining the effects of anthropoge nise on marine mammals is currently an active area of research. The developemnt methods to detect and localixe the animals in their native environments is key to advance this reearch and our understanding .This paper presents a collection of algorithms for automated passive acoustic detection, classification and localization of vocalizing marine mammals in open ocean environments. The toool set known as M3R (Marine Mammal Monitoring on Navy Ranges) uses the large fields of wide-bandwidth bottom-mounted hydrophones that are part of the U.S Navy undersea ranges to listen for vocalizing whales. M3R employs time-frequency analysis to passively detect whale vocalizations;It then aligns detections among neighboring hydrophones to determine the difference in times of arrival (TDOA) of each vocalization. Sets of TDOA are then used to determine 2-D or 3-D position points using hyperbolic localization techniques. An M3R system is capable of continuous automated real-time monitoring of over 200 wide-bandwidth hydrophones covering 2,000+ km(2) of open ocean M3R is typically used in a collaborative fashion to localize animals in support of tagging exercises visual surveys and behavioural response experiments.

关键词： passive acoustic monitoring marine mammal monitoring detection and localization

来源：评论

学校读者我要写书评

暂无评论

Optimization of Pinhole Imaging System Using Ideal Observer for detection and localization

Optimization of Pinhole Imaging System Using Ideal Observer ...

引用

2007 IEEE Nuclear Science Symposium Conference, vol.10

作者： Lili Zhou Parmeshwar Khurd Santosh Kulkarni Anand Rangarajan Gene Gindi Department of Radiology Stony Brook University Stony Brook NY USA Philips Medical Systems Highland Heights OH USA Department of Electrical & Computer Engineering Stony Brook University Stony Brook NY USA Department of C.I.S.E. University of Florida Gainesville FL USA Department of Electrical & Computer Engineering and Radiology Stony Brook University Stony Brook NY USA

For the familiar 2-class detection problem (signal present/absent), ideal observers have been applied to optimization of pinhole and collimator parameters in planar emission imaging. Given photon noise and background and signal variability, such experiments show how to optimize an aperture to maximize detectability of the signal. Here, we consider a fundamentally different, more realistic task in which the observer is required to both detect and localize a signal. The signal is embedded in a variable background and is known except for location. We inquire whether the addition of a localization requirement changes conclusions on aperture optimization. We have previously formulated an ideal observer for this joint detection/localization task, and here apply it to the classic problem of determining an optimal pinhole diameter in a planar emission imaging system. We conclude that as search tolerance on localization decreases, the optimal pinhole diameter shrinks from that required by detection alone, and in addition, task performance becomes more sensitive to fluctuations about the optimal pinhole diameter. As in the case for detection only, the optimal pinhole diameter shrinks as the amount of background variability grows, and in addition, conspicuity limits can be observed. Unlike the case for detection only, our task leads to a finite aperture size in the absence of background variability. For both tasks, the inclusion of background variability yields a finite aperture size.

关键词： Ideal observer LROC ROC Planar emission imaging Pinhole detection and localization

来源：评论

学校读者我要写书评

暂无评论

An Approach to System Optimization for X-Ray Photon-Counting Systems Using Performance on a detection/localization Task

An Approach to System Optimization for X-Ray Photon-Counting...

引用

60th IEEE Nuclear Science Symposium (NSS) / Medical Imaging Conference (MIC) / 20th International Workshop on Room-Temperature Semiconductor X-ray and Gamma-ray Detectors

作者： Lu, Yihuan Zhang, Hao Liang, Zhengrong Gindi, Gene SUNY Stony Brook Dept Elect & Comp Engn Stony Brook NY 11794 USA

ISBN: (纸本)9781479905348

We address the problem of optimizing data acquisition for photon-counting CT. We formulate a task-driven approach using a clinically relevant task of detection and localization of a lesion in a search region. The appropriate scalar measure of task performance is ALROC, the area under the LROC curve. For hardware optimization, the observer performing the task should operate on the raw (sinogram) data so that the best possible information, independent of the parameters of any particular reconstruction algorithm, is collected. To carry out the task, we use the ideal observer (IO), a numerical observer that yields the maximal ALROC amongst all observers. This differs from the usual numerical observer that operates on the reconstructed data and is designed so that its performance tracks that of a human observer. So while the optimization is carried out in the sinogram domain, the task definition is in the object domain. In this work, we mathematically specialize a previously developed 10 to the case of photon-counting transmission tomography. We applied the 10 to a 2D simulation of CT with the task of detecting a 3mm lung nodule in a lung region of a 512x512 phantom. The application was to optimize the number of angular acquisitions given a fixed dose. A plot of ALROC vs. angle number peaks at 42 angles. We also plotted ALROC vs. dose (at 105 angles) to characterize the increase in task performance with count level.

关键词： X-ray detection and localization ideal observer system optimization photon-counting

来源：评论

学校读者我要写书评

暂无评论

Multi-Scale Proposal Regions Fusion Network for detection and 3D localization of the Infected Trees 3

Multi-Scale Proposal Regions Fusion Network for Detection an...

引用

3rd International Conference on Artificial Intelligence in Information and Communication (IEEE ICAIIC)

作者： Hou, Junling Li, Weihong Gong, Weiguo Wang, Zixu Chongqing Univ Key Lab Optoelect Technol & Syst Educ Minist Chongqing Peoples R China

ISBN: (纸本)9781728176383

Forest surveillance towers have the advantages of long observation time, wide observation range, stable and real-time observation. In this paper, a multi-scale proposal regions fusion network (MFRPN) is proposed for detecting the infected trees automatically on the enhanced images from the forest surveillance towers, which can solve the problem that small and large targets can't be effectively detected on a single scale. The proposed MFRPN includes multi-scale images, three CNNs, three different RPNs, and proposal regions fusion model. In the proposed method, we train and run the scale-specific detectors in a multi-task fashion. And, to obtain the accurate spatial level location information of the infected trees, we achieve the three-dimensional (3D) coordinates localization of the digital elevation model (DEM) by using the principle of forest surveillance towers imaging and terrain elevation data. The experimental results show the detection accuracy achieves 91.63%, the detection time of a single image is 0.46 second, and the 3D localization error is less than 50m. The proposed network can realize the real-time detection and 3D localization of the infected trees.

关键词： multi-scale detection and localization the infected trees deep learning

来源：评论

学校读者我要写书评

暂无评论

Real-time social distance monitoring and face mask detection based Social-Scaled-YOLOv4, DeepSORT and DSFD&MobileNetv2 for COVID-19

引用

MULTIMEDIA TOOLS AND APPLICATIONS 2024年第10期83卷 30613-30639页

作者： Mokeddem, Mohammed Lakhdar Belahcene, Mebarka Bourennane, Salah M Khider Univ RB IAIM LI3C Biskra Algeria Ecole Cent GSM Fresnel Inst Marseille France

COVID-19 has spread rapidly worldwide, despite the availability of vaccines, the fear of the World Health Organization continues due to the mutation of the Coronavirus. This is what prompted us to propose this work of social distance and wearing a face mask to fight against this pandemic to save lives. In this work, we propose a real-time four-stage model with monocular camera and deep learning based framework for automating the task of monitoring social distancing and face mask detection using video sequences. This work based on Scaled-You Only Look Once (Scaled-YOLOv4) object detection model, Simple Online and Real-time Tracking with a deep association metric approach to tracking people. The perspective transformation is used to approximate the three-dimensional coordinates with Euclidean metric to compute distance between boxes. The Dual Shot Face Detector (DSFD) and MobileNetv2 face mask model used to detect faces of people who violate or cross the social distance. Accuracy of 56.2% and real-time performance of 32 frames per second are achieved by the Social-Scaled-YOLOv4 (Social-YOLOv4-P6) model trained on the MS COCO dataset and Google-Open-Image dataset. The results are compared with other popular state-of-the-art models in terms of Mean-Average-Precision, frame rate and loss of values. The DSFD&MobileNetv2 facemask detectors trained on Wider Face and Real Face mask dataset achieves an accuracy of 99.3%. The proposed approach is validated on indoor/outdoor public images and video sequences such as wider face dataset, Oxford Town Center dataset and open access sequences.

关键词： COVID-19 detection and localization Deep Learning Social Distancing Scaled-YOLOv4 Tracking

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：