检索结果-内蒙古大学图书馆

HFR-YOLO v3 model for multi-object detection and unscented Kalman filtering based object tracking in vehicle driving footage

引用

INTERNATIONAL JOURNAL OF INTELLIGENT ROBOTICS AND APPLICATIONS 2025年 1-19页

作者： Manickaraj, Dhanalakshmi Mann, Palvinder Singh Gujarat Technol Univ Sch Engn & Technol Ahmadabad Gujarat India

multi-object detection and tracking in vehicle footage supports a lot of applications, such as video analytics, smart transportation systems, video surveillance, and robot navigators. Many existing approaches in object tracking and detection result in poor detection performance. To tackle this challenge, this research proposes a Hybrid Fast Regional-You Only Look Once v3 (HFR-YOLO v3) technique for multi-object detection and unscented Kalman filtering-based object tracking in vehicle driving footage. The proposed technique is implemented as follows. The input video is sourced from the dataset and passed through the frame extraction module, which divides the video into individual frames for subsequent processing. Then, the object is segmented from the frame, which is performed by the Entropy Weighting K-means Algorithm. After that, multiple object detection is performed by creating an anchor box using HFR-YOLO v3, which is developed by the amalgamation of YOLO v3 and Fast Regional Convolution Neural Network (FR-CNN). Finally, the multi-object tracking is performed, which is done by Kalman filtering by modifying the moving model. Moreover, the HFR-YOLO v3 accomplished a higher true positive rate (TPR) of 0.977, true negative rate (TNR) of 0.948, and accuracy of 0.958.

关键词： multi-object detection Fast Regional Convolution Neural Network Unscented Kalman filtering YOLO v3 multi-object tracking

来源：评论

学校读者我要写书评

暂无评论

multi-object detection for Daily Road Maintenance Inspection With UAV Based on Improved YOLOv8

引用

IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS 2024年第11期25卷 16548-16560页

作者： Zhu, Junqing Wu, Yuxuan Ma, Tao Southeast Univ Sch Transportat Nanjing 211189 Peoples R China

Daily road maintenance is essential to road safety and serviceability, particularly to highways. In daily road maintenance inspection tasks, the objectives include a variety of targets, such as pavement cracks, foreign objects, guardrail damages etc. There is a lack of rapid detection methods that allow for the uniform identification of multiple targets for road maintenance. This paper proposes an automatic multi-object detection method for road daily maintenance inspection assisted by unmanned aerial vehicles ( UAV). A dataset of multiple roadway anomalies ( UAVROAD) for daily road maintenance needs was constructed. UM-YOLO, an improved algorithm based on the YOLOv8n algorithm was created to better extract the features of multiple targets, as well as fuse features and reduce computation while maintaining accuracy. The improvements include adding EMA (Efficient multiscale Attention Module) in the C2f module in the backbone, employing Bi-FPN fusion mechanism in the neck and using GSConv, a lightweight convolutional network, for the convolution operation. Compared with the YOLOv8n, the proposed UM-YOLO improved mean average precision(mAP) by 4.6% and reduced the model computation by 14%.

关键词： Roads Maintenance Inspection Autonomous aerial vehicles Feature extraction Task analysis Accuracy Daily road maintenance UAV multi-object detection computer vision deep learning

来源：评论

学校读者我要写书评

暂无评论

multi-object detection of Chinese License Plate in Complex Scenes

引用

Computer Systems Science & Engineering 2021年第1期36卷 145-156页

作者： Dan Liu Yajuan Wu Yuxin He Lu Qin Bochuan Zheng School of Computer Science China West Normal UniversityNanchong637002China School of Mathematics and Information China West Normal UniversityNanchong637002China Internet of Things Perception and Big Data Analysis Key Laboratory of Nanchong China West Normal UniversityNanchong637002China

multi-license plate detection in complex scenes is still a challenging task because of multiple vehicle license plates with different sizes and classes in the images having complex *** edge features of high-density distribution and the high curvature features of stroke turning of Chinese character are important signs to distinguish Chinese license plate from other *** accurately detect multiple vehicle license plates with different sizes and classes in complex scenes,a multi-object detection of Chinese license plate method based on improved YOLOv3 network was proposed in this *** improvements include replacing the residual block of the YOLOv3 backbone network with the Inception-ResNet-A block,imbedding the SPP block into the detection network,cutting the redundant Inception-ResNet-A block to suit for the multi-license plate detection task,and clustering the ground truth boxes of license plates to obtain a new set of anchor boxes.A Chinese vehicle license plate image dataset was built for training and testing the improved network,and the location and class of the license plates in each image were accurately *** dataset has 62,153 pieces of images and 4 classes of China vehicle license plates,almost images have multiple license plates with different *** demonstrated that the multilicense plate detection method obtained 83.4%mAP,98.88%precision,98.17%recall,98.52 F1 score,89.196 BFLOPS and 22 FPS on the test dataset,and whole performance was better than the other five compared networks including YOLOv3,SSD,Faster-RCNN,EfficientDet and RetinaNet.

关键词： Chinese vehicle license plate multiple license plate multi-object detection Inception-ResNet-A spatial pyramid pooling

来源：评论

学校读者我要写书评

暂无评论

multi-object detection Method based on YOLO and ResNet Hybrid Networks 4

Multi-object Detection Method based on YOLO and ResNet Hybri...

引用

4th IEEE International Conference on Advanced Robotics and Mechatronics (ICARM)

作者： Lu, Zhenyu Lu, Jia Ge, Quanbo Than, Tianming Nanjing Univ Informat Sci & Technol Sch Elect & Informat Engn Nanjing 210044 Jiangsu Peoples R China Shenzhen Inst Guangdong Ocean Univ Shenzhen 518120 Peoples R China Hangzhou Dianzi Univ Sch Automat Inst Syst Sci & Control Engn Hangzhou 310018 Peoples R China Nanjing Audit Univ Sch Informat Engn Nanjing 211815 Jiangsu Peoples R China

ISBN: (纸本)9781728100647

As a sub-discipline related to artificial intelligence, machine vision has become a hotspot in recent years. object detection, as part of machine vision, has been widely used in various fields. However, the existing object detection methods based on deep learning theory cannot accurately recognize multiple objects. Thus, in this paper we propose a new hybrid network based on YOLO and ResNet (Yolo-resnet) for multi-object detection. Our method integrates ResNet into the feature extraction of the YOLOv3 framework and the detection results demonstrate that our hybrid network is efficient for detecting multi-objects from an image of complex natural scenes.

关键词： Machine vision Deep learning YOLOv3 Residual network multi-object detection

来源：评论

学校读者我要写书评

暂无评论

multi-object detection Based on Deep Learning in Real Classrooms 15th

Multi-object Detection Based on Deep Learning in Real Classr...

引用

15th Pacific Rim International Conference on Artificial Intelligence (PRICAI) / 15th Pacific Rim Knowledge Acquisition Workshop (PKAW)

作者： Shao, Benchi Jiang, Fei Shen, Ruimin Shanghai Jiao Tong Univ Dept Comp Sci & Engn 800 Dongchuan Rd Shanghai Peoples R China

ISBN: (纸本)9783319973104;9783319973098

Information-based classrooms with cameras provide numerous videos for evaluating teaching qualities. In this paper, we focus on automatically detecting the learning-and-teaching related behaviors for further teaching analysis, including standing-up and hand-raising of students, and movements of teachers. First, due to the continuity of behaviors, we convert it into a frame-based object detection problem. Compared with the publicly-available datasets of object detection, there are several challenges in our real classrooms, such as low resolutions, various gestures, complex backgrounds, and occlusions. Second, to solve these challenges, we propose an improved R-FCN architecture, which incorporates Feature Pyramid Networks into Region Proposal Networks (RPNs) and introduces a position-sensitive RoIAlign layer. The multi-level features and RPNs provide more contexture information for small object detections, and the position-sensitive RoIAlign layer reduces the misalignment in extracting features for region proposals. Lastly, the efficiency of the proposed algorithm is demonstrated on our collected data from real classrooms, which contains 30k frames with 100k labeled objects.

关键词： multi-object detection Deep learning Classroom

来源：评论

学校读者我要写书评

暂无评论

multi-object detection and tracking in video sequences

Multi-object detection and tracking in video sequences

引用

作者： Ala Mhalla UNIVERSITY of Clermont Auvergne - Clermont Ferrand

学位级别：博士

The work developed in this PhD thesis is focused on video sequence analysis. The latter consists of object detection, categorization and tracking. The development of reliable solutions for the analysis of video sequences opens new horizons for several applications such as intelligent transport systems, video surveillance and robotics. In this thesis, we put forward several contributions to deal with the problems of detecting and tracking multi-objects on video sequences. The proposed frameworks are based on deep learning networks and transfer learning approaches. In a first contribution, we tackle the problem of multi-object detection by putting forward a new transfer learning framework based on the formalism and the theory of a Sequential Monte Carlo (SMC) filter to automatically specialize a Deep Con- volutional Neural Network (DCNN) detector towards a target scene. The suggested specialization framework is used in order to transfer the knowledge from the source and the target domain to the target scene and to estimate the unknown target distri- bution as a specialized dataset composed of samples from the target domain. These samples are selected according to the importance of their weights which reflects the likelihood that they belong to the target distribution. The obtained special- ized dataset allows training a specialized DCNN detector to a target scene without human intervention. In a second contribution, we propose an original multi-object tracking frame- work based on spatio-temporal strategies (interlacing/inverse interlacing) and an interlaced deep detector, which improves the performances of tracking-by-detection algorithms and helps to track objects in complex videos (occlusion, intersection, strong motion). In a third contribution, we provide an embedded system for traffic surveillance, which integrates an extension of the SMC framework so as to improve the detection accuracy in both day and night conditions and to specialize any DCNN detector for bo

关键词： Artificial intelligence Computer vision Transfer learning Deep learning multi-object detection Specialization Tracking-by-detection multi-object tracking Em-bedded system

来源：评论

学校读者我要写书评

暂无评论

Progress in multi-object detection models: a comprehensive survey

引用

multiMEDIA TOOLS AND APPLICATIONS 2023年第15期82卷 22405-22439页

作者： Balakrishna, Sivadi Mustapha, Ahmad Abubakar Vigans Fdn Sci Technol & Res Dept Comp Sci & Engn Guntur Andhra Pradesh India

Deep learning-based object detection has become popular due to its strong learning ability and advantages in dealing with occlusion, scale transformation, and context changes. In recent years, it has become a research hotspot. This paper presents the current Deep Learning models from Generic and Salient detection models ranging from one-stage to two-stage for multi-object detection in various applications. Nevertheless, we also examined the advantages and some drawbacks of those models. Furthermore, challenges such as variation in object scales, computation time, illumination differing from various applications, and promising research directions of Deep Learning models are discussed. Finally, we proposed Dense PRediction Simplified (DPRS) based on the YOLO model. Backbones play a vital role in enhancing the performance of detection models, and efficient Backbone architecture will be fused to achieve the competitive state-of-art result.

关键词： Deep learning multi-object detection Models DPRS Challenges Research directions

来源：评论

学校读者我要写书评

暂无评论

A method of cross-layer fusion multi-object detection and recognition based on improved faster R-CNN model in complex traffic environment ?

引用

PATTERN RECOGNITION LETTERS 2021年 145卷 127-134页

作者： Li, Cui-jin Qu, Zhong Wang, Sheng-ye Liu, Ling Chongqing Univ Posts & Telecommun Coll Comp Sci & Technol Chongqing 400065 Peoples R China Chongqing Univ Posts & Telecommun Coll Mobile Telecommun Chongqing 401520 Peoples R China

Improving the detection accuracy and speed is the prerequisite of multi-object recognition in the complex traffic environment. Despite object detection has made significant advances based on deep neural networks, it remains a challenge to focus on small and occlusion objects. We address this challenge by allowing multiscale fusion. We introduce a cross-layer fusion multi-object detection and recognition algorithm based on Faster R-CNN, an approach that the five-layer structure of VGG16 (Visual Geometry Group) is used to obtain more characteristic information. We implement this idea with lateral embedding the 1 & times;1 convolution kernel, max pooling and deconvolution, in conjunction with weighted balanced multi-class cross entropy loss function and Soft-NMS to control the imbalance between difficult and easy samples. Considering the actual situation in a complex traffic environment, we manually label mixed dataset. On Cityscapes and KITTI datasets, experimental results show that the proposed model achieves better effects than the current mainstream object detection models. (c) 2021 Elsevier B.V. All rights reserved.

关键词： multi-object detection multi-object recognition Faster R-CNN Weighted balanced multi-class cross entropy loss function

来源：评论

学校读者我要写书评

暂无评论

Deep Transfer Learning-Based multi-object detection for Plant Stomata Phenotypic Traits Intelligent Recognition

引用

IEEE-ACM TRANSACTIONS ON COMPUTATIONAL BIOLOGY AND BIOINFORMATICS 2023年第1期20卷 321-329页

作者： Yang, Xiao-Hui Xi, Zi-Jun Li, Jie-Ping Feng, Xin-Lei Zhu, Xiao-Hong Guo, Si-Yi Song, Chun-Peng Henan Univ Sch Math & Stat Henan Engn Res Ctr Artificial Intelligence Theory Kaifeng 475004 Henan Peoples R China Cent South Univ Sch Comp Sci Changsha 410017 Hunan Peoples R China Henan Univ Sch Life Sci Dept Plant Sci Key Lab Plant Stress BiolState Key Lab Cotton Bi Kaifeng 475004 Henan Peoples R China

Plant stomata phenotypic traits can provide a basis for enhancing crop tolerance in adversity. Manually counting the number of stomata and measuring the height and width of stomata obviously cannot satisfy the high-throughput data. How to detect and recognize plant stomata quickly and accurately is the prerequisite and key for studying the physiological characteristics of stomata. In this research, we consider stomata recognition as a multi-object detection problem, and propose an end-to-end framework for intelligent detection and recognition of plant stomata based on feature weights transfer learning and YOLOv4 network. It is easy to operate and greatly facilitates the analysis of stomata phenotypic traits in high-throughput plant epidermal cell images. For different cultivars, multi-scales, rich background features, high density, and small stomata object images, the proposed method can precisely locate multiple stomata in microscope images and automatically give phenotypic traits of stomata. Users can also adjust the corresponding parameters to maximize the accuracy and scalability of automatic stomata detection and recognition. Experimental results on actual data provided by the National Maize Improvement Center show that the proposed method is superior to the existing methods in high stomata automatic detection and recognition accuracy, low training cost, strong generalization ability.

关键词： Deep learning transfer learning multi-object detection stomata recognition

来源：评论

学校读者我要写书评

暂无评论

Moving multi-object detection and Tracking Using MRNN and PS-KM Models

引用

Computer Systems Science & Engineering 2023年第2期44卷 1807-1821页

作者： V.Premanand Dhananjay Kumar Department of Information Technology Anna UniversityMIT CampusChennai600044India

On grounds of the advent of real-time applications,like autonomous driving,visual surveillance,and sports analysis,there is an augmenting focus of attention towards multiple-object Tracking(MOT).The tracking-by-detection paradigm,a commonly utilized approach,connects the existing recognition hypotheses to the formerly assessed object trajectories by comparing the simila-rities of the appearance or the motion between *** an efﬁcient detection and tracking of the numerous objects in a complex environment,a Pearson Simi-larity-centred Kuhn-Munkres(PS-KM)algorithm was proposed in the present *** this light,the input videos were,initially,gathered from the MOT dataset and converted into *** background subtraction occurred whichﬁltered the inappropriate data concerning the frames after the frame conversion ***,the extraction of features from the frames was ***,the higher dimensional features were transformed into lower-dimensional features,and feature reduction process was performed with the aid of Information Gain-centred Singular Value Decomposition(IG-SVD).Next,using the Modiﬁed Recurrent Neural Network(MRNN)method,classiﬁcation was executed which identiﬁed the categories of the objects *** PS-KM algorithm identi-ﬁed that the recognized objects were ***,the experimental outcomes exhibited that numerous targets were precisely tracked by the proposed system with 97%accuracy with a low false positive rate(FPR)of 2.3%.It was also proved that the present techniques ***,CNN,and KNN,were effective with regard to the existing models.

关键词： multi-object detection object tracking feature extraction morlet wavelet mutation(MWM) ant lion optimization(ALO) background subtraction

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：