检索结果-内蒙古大学图书馆

Automatic Data Cleaning System for Large-Scale Location Image Databases Using a Multilevel Extractor and Multiresolution Dissimilarity Calculation

引用

IEEE INTELLIGENT SYSTEMS 2021年第5期36卷 49-56页

作者： Cheng, Hsu-Yung Yu, Chih-Chang Natl Cent Univ Dept Comp Sci & Informat Engn Taoyuan 320 Taiwan Chung Yuan Christian Univ Dept Informat & Comp Engn Taoyuan 320 Taiwan

In this article, we propose a system for automatically classifying and cleaning location images in large-scale image databases uploaded by arbitrary users. Detecting incorrect scenes uploaded by users and maintaining the correctness of the database through automatic data cleaning are essential because human inspection is not feasible for verifying massive amounts of data. In this study, we compared different feature extractors using deep convolutional neural networks trained using big data. We designed a multilevel extractor to improve feature extraction. Moreover, a detector based on multiresolution dissimilarity calculation was designed to overcome the issue of large intraclass distances and successfully identify incorrect scenes. The proposed system was validated using a highly challenging dataset with 138,000 images collected from Google Places. The experiments show that the multilevel extractor and the detector based on multiresolution dissimilarity calculation can improve the accuracy in identifying incorrect scenes and achieve satisfying data cleaning results.

关键词： Feature extraction Databases Image resolution Cleaning Google detection algorithms Intelligent systems Automatic data cleaning deep learning location image analysis

来源：评论

学校读者我要写书评

暂无评论

Decoupled Two-Stage Crowd Counting and Beyond

引用

IEEE TRANSACTIONS ON IMAGE PROCESSING 2021年 30卷 2862-2875页

作者： Cheng, Jian Xiong, Haipeng Cao, Zhiguo Lu, Hao Huazhong Univ Sci & Technol Key Lab Image Proc & Intelligent Control Minist Educ Wuhan 430074 Peoples R China Huazhong Univ Sci & Technol Sch Artificial Intelligence & Automat Wuhan 430074 Peoples R China

One of appealing approaches to counting dense objects, such as crowd, is density map estimation. Density maps, however, present ambiguous appearance cues in congested scenes, rendering infeasibility in identifying individuals and difficulties in diagnosing errors. Inspired by an observation that counting can be interpreted as a two-stage process, i.e., identifying possible object regions and counting exact object numbers, we introduce a probabilistic intermediate representation termed the probability map that depicts the probability of each pixel being an object. This representation allows us to decouple counting into probability map regression (PMR) and count map regression (CMR). We therefore propose a novel decoupled two-stage counting (D2C) framework that sequentially regresses the probability map and learns a counter conditioned on the probability map. Given the probability map and the count map, a peak point detection algorithm is derived to localize each object with a point under the guidance of local counts. An advantage of D2C is that the counter can be learned reliably with additional synthesized probability maps. This addresses important data deficiency and sample imbalanced problems in counting. Our framework also enables easy diagnoses and analyses of error patterns. For instance, we find that, the counter per se is sufficiently accurate, while the bottleneck appears to be PMR. We further instantiate a network D2CNet in our framework and report state-of-the-art counting and localization performance across 6 crowd counting benchmarks. Since the probability map is a representation independent of visual appearance, D2CNet also exhibits remarkable cross-dataset transferability. Code and pretrained models are made available at: https://***/d2cnet

关键词： Location awareness Probabilistic logic Training Reliability Object recognition Kernel detection algorithms Crowd counting crowd analysis convolutional neural networks local count models

来源：评论

学校读者我要写书评

暂无评论

High Dynamic Range Image Recovery by Use of Lens Flare Events detection Algorithm

High Dynamic Range Image Recovery by Use of Lens Flare Event...

引用

International Conference on Electronics, Information and Communications (ICEIC)

作者： Bobaro Chang Hyunsurk Ryu Hyuk-Jae Lee Department of Electrical and Computer Engineering Seoul National University Seoul Korea

Event cameras are novel vision sensors that asynchronously output per-pixel events by measuring brightness changes. Event cameras have advantages such as high speed and high dynamic range compared to conventional cameras. To leverage the advantages and apply prior algorithms, event-based image reconstruction has been developed. With the development of neural networks, state-of-the-art reconstruction methods are introduced. However, high dynamic range reconstructions still suffer from lens flare-based artefacts, which makes the intensity estimation incorrect. To address this problem, this work presents a computational approach to detect ill-posed events caused by lens flare. Under the examination that lens flare elements of event cameras mainly consist of glow and starburst, we derive two bivariate Gaussian distributions from targets in the compressed stream of events. By operating convolution, the detector reduces the sparsity of events, which makes the surface fitting more precise. We show that the proposed method effectively eliminates dark stain in high dynamic range reconstruction, while preserving detail on the other region at the same time.

关键词： Surface reconstruction Three-dimensional displays Heuristic algorithms Cameras High dynamic range detection algorithms Voltage control

来源：评论

学校读者我要写书评

暂无评论

Ellipse Encoding for Arbitrary-Oriented SAR Ship detection Based on Dynamic Key Points

引用

IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING 2022年 60卷 1页

作者： Gao, Fei Huo, Yiyang Sun, Jinping Yu, Tao Hussain, Amir Zhou, Huiyu Beihang Univ Sch Elect & Informat Engn Beijing 100191 Peoples R China Taiyuan Satellite Launch Ctr Xinzhou 36300 Peoples R China Edinburgh Napier Univ Cyber & Big Data Res Lab Edinburgh EH11 4BN Midlothian Scotland Univ Leicester Sch Comp & Math Sci Leicester LE1 7RH Leics England

In recent years, there has been growing interest in developing oriented bounding box (OBB)-based deep learning approaches to detect arbitrary-oriented ship targets in synthetic aperture radar (SAR) images. However, most existing OBB-based detection methods suffer from boundary discontinuity problems for bounding box angle prediction and key point regression challenges. In this article, we present a novel OBB-based detection algorithm that utilizes ellipse encoding to effectively exploit the geometric and scattering properties of ship targets. Specifically, the ship contour is fit by an OBB inscribed ellipse that is encoded as a set of distances between dynamic key points on the bow and target center. By combining the bow angle interval and the decoding process, the negative impact of the boundary discontinuity problem is avoided. In addition, we propose an elliptical Gaussian distribution heatmap and a pooling strategy termed double-peak max-pooling (DPM) to deal with the challenge of separating densely distributed ships in inshore scenes. The former can enhance the heatmap's ship-side score gap between neighboring ship targets, while the latter can solve the problem of target center responses being suppressed after max-pooling. Simulation experiments conducted on the benchmark Rotating SAR Ship detection Dataset (RSSDD) and Rotated Ship detection Dataset in SAR Images (RSDD-SAR) demonstrate the superior performance of our method for ship target detection compared to several state-of-the-art OBB-based algorithms. Ablation experiments show that elliptical Gaussian distribution heatmap and DPM can further improve the inshore detection performance.

关键词： Marine vehicles Feature extraction detection algorithms Synthetic aperture radar Radar polarimetry Heating systems Scattering Arbitrary-oriented boundary discontinuity ellipse encoding ship detection synthetic aperture radar (SAR)

来源：评论

学校读者我要写书评

暂无评论

A novel high-through cell micronucleus detection algorithm based on improved YOLOv5

A novel high-through cell micronucleus detection algorithm b...

引用

Industrial IoT, Big Data and Supply Chain (IIoTBDSC), International Conference on

作者： Weiyi Wei Jingyu Li Xiaoqin Wu Hangjian Zhang College of Computer Science and Engineering Northwest Normal University Lanzhou China Radiology Department Gansu Provincial Center For Disease Control And Prevention Lanzhou China Business school Northwest Normal University Lanzhou China

ISBN: (数字)9798350341690

ISBN: (纸本)9798350341706

In response to the challenges associated with cell micronucleus image recognition, namely small target sizes, high class similarity, and limited sample data, A novel cell micronucleus detection algorithm named YOLOv5-ECA is proposed based on YOLOv5. The proposed algorithm gets hyper-accurate micronucleus identification through a series of improvements. To begin with, the anchor frame size is recalibrated based on a meticulously constructed dataset. This process serves to expedite convergence speed while bolstering accuracy. Subsequently, in C3 module, a coordinate attention (CA) is introduced, which enlarging the receptive field and enabling the model to allocate greater focus to crucial elements within processed images. Ultimately, a superficial detection layer is incorporated, thereby enhancing the precision of cell micronucleus identification through multi-scale detection. Comparative analysis with the original YOLOv5 which demonstrates a 2% increase in average precision, accuracy, and a 1.3% enhancement in recall for the YOLOv5-ECA algorithm based on experimental results.

关键词： YOLO Accuracy Heuristic algorithms Supply chains Real-time systems detection algorithms Optimization

来源：评论

学校读者我要写书评

暂无评论

YOLOv5 Water Surface Garbage detection Algorithm Based on Deep Separable Convolution

YOLOv5 Water Surface Garbage Detection Algorithm Based on De...

引用

International Conference on Mechanical and Electronics Engineering (ICMEE)

作者： Xuchun Wu Ping Deng Peng Wang Tianli Ma Mountain River Basin Management Office of kashi area Xin Jiang China Electronic Information Engineering Xi'an Technological University Xi'an China

ISBN: (数字)9798350307146

ISBN: (纸本)9798350307153

The challenges posed by multiple parameters, high computational complexity, and a single convolutional kernel in YOLOv5 lead to subpar accuracy and slow detection rates when applied to water surfaces. To address these issues, we propose a novel YOLOv5 water surface target detection algorithm based on deep separable convolution. This algorithm employs deep separable convolution to reconstruct the Darknet53 network, effectively reducing computational complexity. Additionally, we integrate the Inception network structure into the detection network, elevating both complexity and detection accuracy. Evaluation of the algorithm on the dataset reveals outstanding performance, boasting an impressive 95.3% accuracy and a rapid detection rate of 65.34f/s. These results significantly surpass the original YOLOv5 water surface garbage detection algorithm, demonstrating the efficacy of our proposed method.

关键词： YOLO Training Surface reconstruction Convolution Refining Environmental monitoring detection algorithms

来源：评论

学校读者我要写书评

暂无评论

LOVD: Land Vehicle detection in Complex Scenes of Optical Remote Sensing Image

引用

IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING 2022年 60卷 1页

作者： Yan, Puti Liu, Xinrui Wang, Feng Yue, Chengfei Wang, Xin Harbin Inst Technol Dept Aerosp Engn Harbin 150001 Peoples R China Harbin Inst Technol Shenzhen Inst Space Sci & Appl Technol ISSAT Shenzhen 518055 Peoples R China Tsinghua Univ Dept Comp Sci & Technol Beijing 100084 Peoples R China

Nowadays, there is a growing body of research about object detection in remote sensing data. However, the detection algorithms for small targets in remote sensing areas are inadequate, largely because of the unavailability of high-quality datasets. Most remote sensing datasets are comprehensive, which means that they include bridges, airplanes, and lots of other common categories. Compared with other categories, the number and diversity of weak objects, such as vehicles, are quite insufficient. These limitations greatly affect the detection of small targets in remote sensing images. In order to promote the development of algorithms for the detection of small targets in remote sensing images and also allow access to remote sensing data, we have established a large-scale dataset for the detection of vehicle targets in optical remote sensing images and called it LOVD. It contains 1196 pictures and 541751 instances, covering 13 categories. For the dataset, we have proposed in this article: 1) it is the largest one in terms of vehicle category and the total number of vehicle instances;2) it contains images with various backgrounds in different weather and scenarios;and 3) all targets are marked by oriented bounding boxes (OBBs), and two label formats are provided. Finally, we test the state-of-the-art detection algorithms on our dataset and provide a benchmark for OBB detection.

关键词： Remote sensing Meteorology Training Optical imaging Object detection Detectors detection algorithms Dataset remote sensing vehicle detection

来源：评论

学校读者我要写书评

暂无评论

Add Anchor to the Anchor-Free Object detection Algorithm

Add Anchor to the Anchor-Free Object Detection Algorithm

引用

Artificial Intelligence Technology (ACAIT), Asian Conference on

作者： Heng Song Nan Hu Tianbao Geng Anhui Shuzhi Construction Research Institute Co. Ltd China Railway No.4 Engineering Group Co. Ltd Hefei China

ISBN: (数字)9798350359145

ISBN: (纸本)9798350359152

Anchor free framework is more in line with the principle of visual semantics, but it falls into the problem of visual regression of the target shape (width, height and rotation angle) directly, with low accuracy. In this work, we propose a method to improve the accuracy of anchor-free frames by using anchor techniques. On the basis of object location information estimation, only a few anchor-assisted target shape regressions are generated. The method integrates preset boxes into anchor-free framework skillfully and avoids the huge number of anchors in anchor-based scheme. In addition, we have implemented two detection algorithms for horizontal and rotating frames. Compared with the existing anchor-based and anchor-free schemes, our scheme achieves an optimal balance between detection accuracy and computation. The experiments on the two public data sets, MS COCO and HRS2016, fully showed that the precision of our AP reached state-of-the-art.

关键词： Visualization Costs Shape Semantics Estimation Object detection detection algorithms

来源：评论

学校读者我要写书评

暂无评论

BS-SC: An Unsupervised Approach for Detecting Shilling Profiles in Collaborative Recommender Systems

引用

IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING 2021年第4期33卷 1375-1388页

作者： Cai, Hongyun Zhang, Fuzhi Yanshan Univ Sch Informat Sci & Engn Qinhuangdao 066000 Hebei Peoples R China Hebei Univ Sch Cyber Secur & Comp Baoding 071000 Peoples R China Key Lab Comp Virtual Technol & Syst Integrat Hebe Qinhuangdao 066000 Hebei Peoples R China

Collaborative recommender systems are vulnerable to shilling attacks. To address this issue, many methods including supervised and unsupervised have been proposed. However, supervised detection methods require training classifiers and they only apply to detect known types of attacks. The existing unsupervised detection methods need to know the prior knowledge of attacks, otherwise they suffer from low detection precision. In this paper, we present BS-SC, an unsupervised approach for detecting shilling profiles, which does not need to know the attack size or to label the candidate spammers. BS-SC starts from an in-depth analysis of user behaviors and uses two key mechanisms (i.e., behavior features extraction and behavior similarity matrix clustering) to distinguish shilling profiles from genuine ones. The behavior features reflect the behavior difference between genuine and shilling profiles, and the behavior similarity matrix clustering is to cluster shilling profiles based on their highly similar behaviors. Experimental results on the MovieLens and the sampled Amazon review datasets indicate that BS-SC outperforms the baseline unsupervised approaches, even when the prior knowledge is given for them.

关键词： Feature extraction Recommender systems Collaboration Training Correlation Clustering algorithms detection algorithms Collaborative recommender systems shilling attacks detection behavior analysis spectral clustering

来源：评论

学校读者我要写书评

暂无评论

Community Hiding by Link Perturbation in Social Networks

引用

IEEE TRANSACTIONS ON COMPUTATIONAL SOCIAL SYSTEMS 2021年第3期8卷 704-715页

作者： Chen, Xianyu Jiang, Zhongyuan Li, Hui Ma, Jianfeng Yu, Philip S. Xidian Univ Sch Cyber Engn Xian 710071 Peoples R China Univ Illinois Dept Comp Sci Chicago IL 60607 USA

Complex social network is a kind of relationship system composed of many nodes according to social relations. Community detection helps scholars to understand this network topology and find out meaningful communities. Many scholars are therefore actively exploring new community detection algorithms. However, it brings privacy issues such as the disclosure of personal or group information of community members and goes against an individual or group desire to be hidden. Hence, how to hide a target community in a network to resist the community detection algorithms becomes critical. Given this, this article studies the community hiding problem, which has not been extensively focused so far and aims to properly hide a target community into other communities by perturbing a budget limited number of social links. First, we formalize the community hiding problem and design a gain function. Second, we prove the feasibility of link perturbation operations and propose an efficient algorithm to solve the community hiding problem. Finally, we conduct extensive experiments on varieties of real small- and large-scale social networks and compare our method with other community deception algorithms. The experimental results demonstrate that the proposed algorithm is more efficient than previously proposed algorithms.

关键词： Social networking (online) Perturbation methods detection algorithms Law enforcement Privacy Dispersion Security Community deception community detection community hiding community security link perturbation

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：