检索结果-内蒙古大学图书馆

SSRN 2023年

作者： Wang, Hengyou Song, Kani Jiang, Xiang He, Zhiquan School of Science Beijing University of Civil Engineering and Architecture Beijing100044 China School of Computer Science and Technology Jiangsu Normal University Xuzhou221116 China Guangdong Multimedia Information Service Engineering Technology Shenzhen University Shenzhen518060 China

Image captioning has become one of the most popular research problems in the field of artificial intelligence. Although many studies have achieved excellent results, there still are some challenges, for example, cross-modal feature alignment lacks explicit guidance, and model-generated sentences contain grammatical errors. In this paper, we propose a relationship-aligned and grammar-wise BERT model, which integrates a relationship exploration module and a grammar enhancement module into the BERT-based model. Specifically, in the relationship exploration module, to explore relationship tags as anchors to guide semantic alignment, we design a network to calculate the cosine similarity between visual features and word vector information. For the grammar enhancement module, we construct it similarly to the BERT. That means we use two BERT modules in our framework. The first one is the main frame for generating captions, and the second one is the auxiliary model to determine whether the syntax of generated caption is correct or not. To validate the performance of our proposed model, we conduct abundant experiments on the MSCOCO dataset, Flickr30k dataset, and Flickr8k dataset. Experimental results show that our proposed method has a better performance compared to state-of-the-art approaches. © 2023, The Authors. All rights reserved.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Trajectory Prediction and Multi-Risk-Factors Based Dynamic Collision Prediction Method between Construction Vehicles and Workers in Construction Sites

SSRN

引用

SSRN 2024年

作者： Gan, Wenxia Gu, Kedi Yang, Ruqin Geng, Jing Wang, Huini Qiu, Canzhi Hu, Xiaodi School of Civil Engineering and Architecture Wuhan Institute of Technology Hubei Wuhan430074 China Hubei Land Surveying and Mapping Institute Hubei Wuhan430074 China School of Computer Science & Technology Beijing Institute of Technology Beijing100081 China

Predicting the trajectory of construction vehicles and workers at construction sites is crucial for preventing collision accidents and enhancing safety. However, current prediction methods based on trajectory prediction often focus on the trajectory of workers or construction vehicles and failing to fully exploit critical risk factors, i.e., proximity, velocity, and hazardous zones etc. Therefore, this paper proposes a collision prediction method for construction vehicles and workers by combining trajectory prediction techniques in computer vision with a multiple risk factors-based collision risk assessment. The effectiveness of the method are evaluated across different construction site scenarios. The transformer-based trajectory prediction model demonstrates superior accuracy compared to classical models, such as social-GAN and social-LSTM, which the average displacement error is 0.53m construction site. The research findings demonstrate that proposed collision prediction method can be used to accurately predicting collision risk level of workers on construction sites. © 2024, The Authors. All rights reserved.

关键词： Trajectories

来源：评论

学校读者我要写书评

暂无评论

Keyword-fusion Pointer-Generator Network for Policy Text Title Summarization 4

Keyword-fusion Pointer-Generator Network for Policy Text Tit...

引用

4th IEEE International Conference on Information Systems and computer Aided Education, ICISCAE 2021

作者： Gu, Ziyin Chen, Li Zhu, Qingmeng Li, Lingbo Zhang, Zelin Zhou, Xin University of Chinese Academy of Sciences College of Computer Science and Technology Beijing China School of Civil Engineering and Architecture Northeast Electric Power University Jilin China

ISBN: (纸本)9781665441247

In this paper, we study the electricity price policy text title summarization problem. Comparing with conventional summarization tasks, title summarization of policy text has an extra characteristic. Policy texts always contain many professional keywords. In order to retain the main information in title summarization as much as possible, we propose keyword-fusion pointer-generator network with additional consideration of keywords of policy text. We incorporate keywords information from the original policy texts into our model by a new attention mechanism called keyword-fusion attention mechanism so that keywords can be generated in the title. What's more, our keyword-fusion pointer-generator network contains a more useful coverage vector using exponentially weighted averages method in order to solve the problem of repetition. Experimental results show that our model outperforms the other baselines. © 2021 IEEE.

关键词： Adaptation models Conferences Computational modeling Task analysis Information systems

来源：评论

学校读者我要写书评

暂无评论

Neuromorphic Synergy for Video Binarization

arXiv

引用

arXiv 2024年

作者： Lin, Shijie Zhang, Xiang Yang, Lei Yu, Lei Zhou, Bin Luo, Xiaowei Wang, Wenping Pan, Jia The Department of Computer Science The University of Hong Kong Centre for Transformative Garment Production Hong Kong The Centre for Transformative Garment Production Hong Kong The School of Electronic and Information Wuhan University Wuhan China The School of Computer Science and Engineering Beihang University Beijing China The Department of Architecture and Civil Engineering City University of Hong Kong Hong Kong The Department of Computer Science and Engineering Texas A&M University TX United States

Bimodal objects, such as the checkerboard pattern used in camera calibration, markers for object tracking, and text on road signs, to name a few, are prevalent in our daily lives and serve as a visual form to embed information that can be easily recognized by vision systems. While binarization from intensity images is crucial for extracting the embedded information in the bimodal objects, few previous works consider the task of binarization of blurry images due to the relative motion between the vision sensor and the environment. The blurry images can result in a loss in the binarization quality and thus degrade the downstream applications where the vision system is in motion. Recently, neuromorphic cameras offer new capabilities for alleviating motion blur, but it is non-trivial to first deblur and then binarize the images in a real-time manner. In this work, we propose an event-based binary reconstruction method that leverages the prior knowledge of the bimodal target’s properties to perform inference independently in both event space and image space and merge the results from both domains to generate a sharp binary image. We also develop an efficient integration method to propagate this binary image to high frame rate binary video. Finally, we develop a novel method to naturally fuse events and images for unsupervised threshold identification. The proposed method is evaluated in publicly available and our collected data sequence, and shows the proposed method can outperform the SOTA methods to generate high frame rate binary video in real-time on CPU-only devices. © 2024, CC0.

关键词： Binary images

来源：评论

学校读者我要写书评

暂无评论

Support vector machine regression(SVR)-based nonlinear modeling of radiometric transforming relation for the coarse-resolution data-referenced relative radiometric normalization(RRN)

引用

Geo-Spatial Information science 2020年第3期23卷 237-247,I0004页

作者： Jing Geng Wenxia Gan Jinying Xu Ruqin Yang Shuliang Wang School of Computer Science and Technology Beijing Institute of TechnologyBeijingChina Beijing Institute of Technology BeijingChina School of Civil Engineering and Architecture Wuhan Institute of TechnologyWuhanChina Wuhan Natural Resources and Planning Information Center WuhanChina

Radiometric normalization,as an essential step for multi-source and multi-temporal data processing,has received critical *** Radiometric Normalization(RRN)method has been primarily used for eliminating the radiometric *** radiometric trans-forming relation between the subject image and the reference image is an essential aspect of *** at accurate radiometric transforming relation modeling,the learning-based nonlinear regression method,Support Vector machine Regression(SVR)is used for fitting the complicated radiometric transforming relation for the coarse-resolution data-referenced *** evaluate the effectiveness of the proposed method,a series of experiments are performed,including two synthetic data experiments and one real data *** the proposed method is compared with other methods that use linear regression,Artificial Neural Network(ANN)or Random Forest(RF)for radiometric transforming relation *** results show that the proposed method performs well on fitting the radiometric transforming relation and could enhance the RRN performance.

关键词： Support Vector machine Regression(SVR) non-linear radiometric transforming relation Relative Radiometric Normalization(RRN) multi-source data

来源：评论

学校读者我要写书评

暂无评论

STRUCTURE-ORIENTED PROGRESSIVE LOW-RANK IMAGE RESTORATION FOR DEFENDING ADVERSARIAL ATTACKS

STRUCTURE-ORIENTED PROGRESSIVE LOW-RANK IMAGE RESTORATION FO...

引用

2021 IEEE International Conference on Multimedia and Expo, ICME 2021

作者： Zhao, Zhiqun Wang, Hengyou Sun, Hao Cao, Wenming He, Zhihai Department of Electrical Engineering and Computer Science University of Missouri ColumbiaMO65211 United States School of Science Beijing University of Civil Engineering and Architecture Beijing100044 China Guangdong Multimedia Information Service Engineering Technology Research Center Shenzhen University Shenzhen518060 China

ISBN: (纸本)9781665438643

Deep neural networks recognize objects by analyzing local image details and summarizing their information along the inference layers to derive the final decision. Because of this, they are prone to adversarial attacks. On the other hand, human eyes recognize objects based on their global structures and semantic cues, instead of local image textures. In this work, we propose to develop a structure-oriented progressive low-rank image completion method to remove unneeded texture details from the input images and shift the bias of deep neural networks towards global object structures and semantic cues. We formulate the problem into a low-rank matrix completion problem with progressively smoothed rank functions to avoid local minimums. Our experimental results demonstrate the proposed method is able to successfully remove the insignificant local image details while preserving important global object structures. © 2021 IEEE computer Society. All rights reserved.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

A Pose and Shape-Aware Cross-Skeleton Motion Retargeting Framework

A Pose and Shape-Aware Cross-Skeleton Motion Retargeting Fra...

引用

IEEE International Conference on Big Data

作者： Chengyu Liu Bo Zhang Wendong Wang Beijing Univ. of Posts and Telecom Beijing China State Key Laboratory of Networking and Switching Technology School of Computer Science (National Pilot Software Engineering School) Beijing University of Posts and Telecommunications

Motion retargeting technology plays a vital role in fields such as computer animation, virtual reality, and gaming industries. Users can save a lot of costs in art design and animation production. However, the current motion retargeting methods still has many strict requirements. For example, some methods require the source and target skeletons need to have the same number of joints or share the same topology. In addition, we find many methods currently use the Mixamo dataset as the training set and test set, and use joint position errors optimize retargeting results, but ignore the source-target differences at the shape geometry level. This may result in interpenetration or loss of contact. Therefore, we introduce a novel framework and use evaluation metrics such as position error related to skeleton and interpenetration ratio related to shape, which makes the retargeting result more realistic.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Flexural Behavior of Basalt Textile-Reinforced Engineered Cementitious Composite Plates after Exposure to Elevated Temperatures

引用

Journal of Composites for Construction 2025年第4期29卷

作者： Yuan, Wanying Liao, Weizhang Li, Huifang Wang, Ying Ma, Chao School of Civil and Transportation Engineering Beijing Univ. of Civil Engineering and Architecture Beijing100044 China

The present work aims to study the residual flexural behavior of basalt textile-reinforced engineered cementitious composite (ECC) systems after exposure to elevated temperatures. A total of 48 plates were prepared and subjected to temperature conditions at 20°C, 200°C, 400°C, and 600°C for 2 h. Then, the samples were tested under four-point bending. The residual mechanical properties (including flexural strength, deflection, and stiffness) and mass loss were examined. A scanning electron microscope was employed to observe microstructural changes in basalt textile-reinforced ECC samples before and after exposure to high temperatures. The results showed that basalt textile-reinforced ECC samples exhibited excellent flexural properties in terms of load-carrying capacity, deformation, energy dissipation capacity, and damage tolerance at temperatures below 200°C. However, at 400°C and 600°C, because of the melting of the polyvinyl alcohol fibers and the decomposition of the hydration products, the stiffness and flexural strength decreased significantly, leading to a brittle failure mode. A simplified theoretical model was proposed for estimating the ultimate moment of the basalt textile-reinforced ECC thin plates, and the calculated values showed a good agreement with the experiments. © 2025 American Society of civil Engineers.

关键词： Bending strength

来源：评论

学校读者我要写书评

暂无评论

A new method for predicting plant proteins function based on multi label classification algorithm

A new method for predicting plant proteins function based on...

引用

IEEE International Conference on Bioinformatics and Biomedicine (BIBM)

作者： Shuai Chen Juan Wang Maozu Guo School of Computer Science Inner Mongolia University Hohhot China School of Electrical and Information Engineering Beijing Key Laboratory of Intelligent Processing for Building Big Data Beijing University of Civil Engineering and Architecture Beijing China

ISBN: (数字)9781665468190

ISBN: (纸本)9781665468206

Protein function annotation is an important content of bioinformatics. It is unrealistic to experimentally annotate a large number of protein functions, so automated prediction of the functions of proteins is required. Studying plant proteins can help us cultivate new plant varieties. However, the existing methods mainly aim at predicting the single function of plant proteins. Most proteins only have sequence information. Therefore, the multi-function prediction method of plant protein based on sequence has become the focus of research. In this study, we propose a method based on sequence to predict the function of plant proteins, called PlantGO. We describe the functional annotation problem as a multi-label classification problem using Gene Ontology (GO) terminology, and predict multiple functions for plant proteins. PlantGO extracts features from three aspects and performs feature fusion respectively. To avoid redundant information in features, PlantGO uses the random forest to select features and obtains three groups of optimal features. PlantGO uses a multi-label learning algorithm MLKNN to predict protein function, which proves the effectiveness and accuracy of the algorithm. Finally, PlantGO adopts the ensemble learning method to integrate all models into a unified model, further improving the prediction performance. Therefore, PlantGO has achieved excellent performance. The performance metrics are Accuracy (0.936), F1 Score (0.944), Ranking loss (0.011), which is better than the previous method in the independent test. We have developed an online server to predict the functions of plant proteins.

关键词： Proteins Protein engineering Annotations Terminology Predictive models Prediction algorithms Feature extraction

来源：评论

学校读者我要写书评

暂无评论

A review of accident prediction in construction: Tools, issues, and solutions 8

A review of accident prediction in construction: Tools, issu...

引用

8th International Symposium on Project Management, ISPM 2020

作者： Guo, Zhonghua Jiang, Hui You, Wan School of Engineering Science University of Chinese Academy of Sciences Beijing100049 China School of Economics and Management Engineering Beijing Univ. of Civil Engineering and Architecture Beijing100044 China

ISBN: (纸本)9781921712722

The frequent occurrence of accidents in construction industry puts forward higher requirements for the prevention of accidents. The key to the prevention of accidents is the accurate prediction of accidents, and corresponding preventive measures are formulated according to the predicted results. The feasibility of accident prediction has been verified in the fields of traffic safety, aviation safety and so on. This paper analyzes the present situation of accident prediction research in construction industry from the aspects of theory, method and content of accident prediction research, and points out the defects of the existing research. The purpose is to make a systematic summary of the existing research for the future research direction. The results of literature analysis show that accurate and effective accident prediction is reflected in three aspects:(1) Objective and true data;(2) Scientific selection of predictive variables and outcome variables;(3) Construction of prediction model. These are the issues that future research should focus on. © 8th International Symposium on Project Management, ISPM 2020. All rights reserved.

关键词： Construction industry

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：