Article
Computer Science, Artificial Intelligence
Jie Mei, Rou-Jing Li, Wang Gao, Ming-Ming Cheng
Summary: This paper proposes a novel connectivity attention network (CoANet) to efficiently extract road information from satellite imagery by leveraging road shapes and connections in graph networks. By utilizing the strip convolution module (SCM) and connectivity attention module (CoA) to capture long-range information and address occlusions in road regions, superior experimental results are achieved.
IEEE TRANSACTIONS ON IMAGE PROCESSING
(2021)
Article
Remote Sensing
Yuan Guo, Bijun Li, Zhi Lu, Jian Zhou
Summary: The paper introduces a novel approach for mining road networks from floating car data, utilizing various algorithms to process and extract road data information, and demonstrating the feasibility and effectiveness of the method through experiments in Wuhan.
GEO-SPATIAL INFORMATION SCIENCE
(2022)
Article
Geochemistry & Geophysics
Zhigang Yang, Daoxiang Zhou, Ying Yang, Jiapeng Zhang, Zehua Chen
Summary: This letter presents an improved encoder-decoder network, RCFSNet, for road extraction from satellite imagery. It incorporates road context and integrates full-stage features to extract complete roads. Experimental results on two public datasets show that RCFSNet outperforms other state-of-the-art methods in terms of road connectivity. Source code is available on GitHub.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2023)
Article
Chemistry, Analytical
Ivan Brkic, Marko Sevrovic, Damir Medak, Mario Miler
Summary: The European Commission has published an EU Road Safety Framework to reduce road fatalities. This study utilized satellite imagery and the Yolo object detector to automatically detect school routes, crosswalks, and divided carriageways. The accuracy achieved for school zones and divided carriageways was 0.988 and 0.950, respectively, while the accuracy for crosswalks ranged from 0.957 to 0.988.
Article
Computer Science, Hardware & Architecture
Vidhi Chaudhary, Preetpal Kaur Buttar, Manoj Kumar Sachan
Summary: Road network is crucial for urban development and various applications. This paper investigates the potential and performance of the U-Net convolution neural network architecture for road detection and achieves improved segmentation results through proposed techniques and hyperparameters.
JOURNAL OF SUPERCOMPUTING
(2022)
Article
Multidisciplinary Sciences
Ethan Brewer, Jason Lin, Peter Kemper, John Hennin, Dan Runfola
Summary: Global investments in road infrastructure are hindered by limited accurate information on existing roads. This study utilizes satellite imagery to estimate road quality and travel speed with promising accuracy. Training a convolutional neural network in the US and fine-tuning it in Nigeria results in successful prediction of road quality in Nigeria.
Article
Engineering, Electrical & Electronic
Jie Wan, Zhong Xie, Yongyang Xu, Siqiong Chen, Qinjun Qiu
Summary: The DA-RoadNet is a road extraction network with dual attention mechanism, designed to effectively solve discontinuous problems and preserve the integrity of extracted roads by utilizing a deep learning network model and a novel attention mechanism module. Additionally, a hybrid loss function is employed to address class imbalance, ensuring stable training of the network model and avoiding local optima.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2021)
Article
Engineering, Civil
Danyang Sun, Fabien Leurent, Xiaoyan Xie
Summary: This study explores individual mobility by analyzing vehicle trajectories and identifying significant places in the Paris region. By using a customized clustering approach, five types of significant places were identified, including home and work places. The results provide insights into human mobility patterns and can assist in urban planning initiatives.
TRANSPORTATION RESEARCH RECORD
(2021)
Article
Computer Science, Artificial Intelligence
Prativa Das, Satish Chand
Summary: This paper proposes a fully convolutional architecture, named Refined DSE-LinkNet, to extract connected and precise road maps by introducing new modules and aggregate loss functions. Experimental results show its superiority on the DeepGlobe dataset, achieving better IoU compared to the winner of DeepGlobe Challenge 2018, D-LinkNet.
CONNECTION SCIENCE
(2021)
Article
Environmental Sciences
Sani Success Ojogbane, Shattri Mansor, Bahareh Kalantar, Zailani Bin Khuzaimah, Helmi Zulhaidi Mohd Shafri, Naonori Ueda
Summary: A novel network based on an end-to-end deep learning framework is proposed for detecting and classifying urban building features, achieving an overall accuracy of over 80%. Morphological operations applied to extracted building footprints have improved the uniformity of building boundaries for increased accuracy in detecting buildings.
Article
Geochemistry & Geophysics
Wei Tian, Xinxin Zhou, Wei Huang, Yonghong Zhang, Pengfei Zhang, Shifeng Hao
Summary: In this study, a novel deep learning model is proposed for estimating tropical cyclone intensity. The model can extract three-dimensional environmental information from multichannel satellite images and uses attention mechanism to focus on core cloud structure and important channels. Experimental results show that the model outperforms traditional methods in terms of accuracy.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2022)
Article
Environmental Sciences
Ziwei Liu, Mingchang Wang, Fengyan Wang, Xue Ji
Summary: Extracting road information from high-resolution remote sensing images is crucial but challenging due to increased spatial heterogeneity between different types of roads. To address this, RALC-Net is proposed to extract a complete and continuous road network using residual attention and local context-aware network, with further enhancement in performance through multi-scale dilated convolution modules.
Article
Computer Science, Artificial Intelligence
Mayank Dixit, Kuldeep Chaurasia, Vipul Kumar Mishra
Summary: This paper introduces a novel deep learning architecture for building extraction from Sentinel-2 satellite images, which outperformed existing models in literature when tested on satellite images from three densely populated cities in India.
EXPERT SYSTEMS WITH APPLICATIONS
(2021)
Article
Chemistry, Multidisciplinary
Lin Gao, Chen Chen
Summary: This study proposes a dual branch dilated pyramid network (DPBFN) for road segmentation in high-resolution remote sensing images. By constructing pyramid features and using multi-head attention blocks, DPBFN is able to extract road features and reduce redundant information. Experimental results demonstrate the stable performance and significant improvement of this method compared to traditional approaches.
APPLIED SCIENCES-BASEL
(2023)
Article
Green & Sustainable Science & Technology
Mia V. Hikuwai, Nicholas Patorniti, Abel S. Vieira, Georgia Frangioudakis Khatib, Rodney A. Stewart
Summary: Artificial Intelligence (AI) is used to detect the presence of asbestos-containing material (ACM) in roofing using advanced algorithms and imagery. The study confirms that combining high-resolution aerial imagery with Mask R-CNN algorithm provides a reliable method for ACM detection, improving the implementation of management policies for ACM.
Article
Computer Science, Artificial Intelligence
Lingbo Liu, Zewei Yang, Guanbin Li, Kuo Wang, Tianshui Chen, Liang Lin
Summary: This paper discusses an important task in land remote-sensing analysis, which is the automatic extraction of traffic roads from remote-sensing data. To address this problem, the authors introduce a novel neural network framework that utilizes the complementary information from different data sources to enhance the accuracy and robustness of road extraction.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Construction & Building Technology
Bing Cui, Huanling Wu, Canhui Zhao, Jiaping Liu, Zhiming Guo
Summary: Nanjing Jiangxinzhou Yangtze River Bridge is an innovative cable-stayed bridge in China with a unique steel-concrete composite structure. With the use of new materials, such as coarse aggregate reactive powder concrete, it has significantly improved the manufacturing speed and performance of the bridge, making significant contributions to the development of cable-stayed bridges.
STRUCTURAL ENGINEERING INTERNATIONAL
(2023)
Article
Construction & Building Technology
Yang Zhou, Yuan Chen, Ming Jin, Jiaping Liu, Changwen Miao
Summary: In this study, molecular dynamics simulations were used to investigate the effects of irradiation on cementitious materials. The results showed that irradiation led to a transformation in structure and a reduction in mechanical performance of calcium silicate hydrates. Additionally, an increase in ductility was observed, which could be beneficial for designing more radiation-resistant cement materials.
JOURNAL OF SUSTAINABLE CEMENT-BASED MATERIALS
(2023)
Article
Materials Science, Paper & Wood
Jianjun Gu, Xiaofei Yan, Dongming Qi, Ruyi Xie, Xiaoming Yang, Yaobang Li, Jiawei Li
Summary: In this study, a durable multifunctional coating was successfully fabricated on cotton fabric, providing it with flame retardant, antibacterial, superhydrophobic and self-cleaning properties. The coated fabric exhibited excellent durability even after multiple washing cycles.
Article
Green & Sustainable Science & Technology
Gang Liu, Zhitao Han, Jiawei Li, Zhen Wang, Shaosi Cheng, Yubin Liu, Jing Yu, Xinxiang Pan
Summary: A novel method for removing pre-oxidized NOx using Na2SO3/THS composite solution is proposed in this study, aiming to achieve high NOx absorption efficiency and low nitrate concentration in scrubbed wastewater simultaneously.
ENVIRONMENTAL PROGRESS & SUSTAINABLE ENERGY
(2023)
Article
Environmental Sciences
Nianping Chi, Jiajun Liu, Minghua Lei, Li Feng
Summary: The introduction of amphiphilic CPAM TP-ADB with cationic microblock structure and hydrophobic association improves the dewatering performance of PD sludge.
ENVIRONMENTAL SCIENCE AND POLLUTION RESEARCH
(2023)
Article
Medical Informatics
Jiajun Liu, Shein-Chung Chow
Summary: In clinical trials, when the primary analysis fails to meet the study objective, the observed positive results in a specific sub-population analysis raise the question of whether they can be used to support regulatory submission. This article proposes several statistical evaluations for confirming the integrity and validity of the observed sub-population analysis results. The evaluations address selection bias, reproducibility, consistency, generalizability, and sensitivity index. A numerical example of a global clinical trial is presented, along with recommendations for incorporating statistical evaluations. Real-world data and evidence are suggested as possible solutions to address regulatory concerns and increase statistical power.
THERAPEUTIC INNOVATION & REGULATORY SCIENCE
(2023)
Article
Oncology
Honghui Zhao, Dan Li, Jiayin Liu, Xinliang Zhou, Jing Han, Long Wang, Zhisong Fan, Li Feng, Jing Zuo, Yudong Wang
Summary: This study analyzed the characteristics of gut microbes in lung cancer patients and found that the differences in gut microbes are related to the effectiveness of anti-PD-1 therapy combined with chemotherapy. In particular, Bifidobacterium breve was identified as a potential biomarker for predicting efficacy.
Article
Computer Science, Interdisciplinary Applications
Maolin Yang, Pingyu Jiang, Jiajun Liu
Summary: Social product design (SPD) is a new pattern that utilizes socialized design resources to enhance design power with lower costs. However, the characteristics of socialized designers (SDs) result in problems such as low collaboration efficiency and trust. To address these problems, a Blockchain and modified Best-worst algorithm enhanced Blackboard model is established. The model improves collaboration efficiency and provides reliable recording and estimation of SDs' contributions during the SPD process, enhancing the operability of SPD application.
COMPUTERS & INDUSTRIAL ENGINEERING
(2023)
Article
Chemistry, Applied
Niuniu Cui, Jiawei Li, Ziwen Xie, Xiaofei Yan, Dongming Qi
Summary: The study introduces a novel pigment dyeing technology that uses charge attractions between cationic nanoscale pigment and cotton fabrics. This technology offers advantages such as no salt, recyclable dyebath, high uptake, and good wet fastness. The research explores the synthesis of cationic copolymer nanoparticle dispersants and the preparation of nanoscale pigment dispersion. The results demonstrate the excellent dispersion performance and colloid stability of the nanoscale pigment dispersion, as well as the higher coloring intensity and color fastness achieved in the dyeing of cotton fabrics. The recyclability of the dyebath significantly reduces dyeing effluent discharge, providing a promising and ecologically sustainable dyeing technology for cotton fabrics.
Article
Polymer Science
Yanning Zeng, Jiawei Li, Chaoying Hu, Bin Yang, Zhao Ning
Summary: Novel self-healing, weldable, and recyclable HTPB-based PU networks with dual reversible covalent bonds were fabricated using one-pot polycondensation. The networks exhibited good mechanical properties and the ability to be recycled.
MACROMOLECULAR CHEMISTRY AND PHYSICS
(2023)
Article
Environmental Sciences
Hongyi Wu, Ling Xie, Yuchen Wu, Liwei Chen, Bian Jiang, Xiaohai Chen, Yinglin Wu
Summary: This study investigated the accumulation of total petroleum hydrocarbons (TPHs) on the coast of Leizhou Peninsula in Southern China. Land-based discharge, sea traffic, and sediment type were found to significantly influence TPHs distribution. The concentration of TPHs increased by 1.036 μg/g with the addition of one more boat on the wharf. 'Blue Carbon' ecosystems, such as mangroves, were more severely polluted. Cleaner production policies should be implemented to mitigate TPHs discharge from coastal areas.
MARINE POLLUTION BULLETIN
(2023)
Article
Spectroscopy
Kang Wang, Gang Li, Mei Zhou, Huiquan Wang, Dan Wang, Ling Lin
Summary: This paper presents a novel method called spectral elimination for noninvasive quantitative analysis of human blood components. The method reduces the influence of non-target components on the detection of target components, resulting in high accuracy. Experimental results demonstrate the effectiveness of the method in predicting the content of multiple human blood components.
SPECTROCHIMICA ACTA PART A-MOLECULAR AND BIOMOLECULAR SPECTROSCOPY
(2023)
Article
Spectroscopy
Qi Xiao, Huajian Luo, Huishan Cao, Bo Li, Jiajia Liu, Yi Liu, Shan Huang
Summary: Ti3C2 quantum dots were synthesized and successfully applied for fluorescent imaging of living cells. The binding interaction between Ti3C2 QDs and trypsin was comprehensively studied using spectroscopic strategies and molecular modeling technique. The results show that Ti3C2 QDs can quench the intrinsic fluorescence of trypsin and significantly inhibit its enzymatic activity on human serum albumin and cells.
SPECTROCHIMICA ACTA PART A-MOLECULAR AND BIOMOLECULAR SPECTROSCOPY
(2023)
Article
Computer Science, Artificial Intelligence
Xiang Shi, Yinpeng Liu, Jiawei Liu, Qikai Cheng, Wei Lu
Summary: Scientific papers are crucial for academic communication, but many of them lack in-depth research and present core content ambiguously, hindering the progression of science and technology. To address this challenge, the INTEGrity vERification (INTEGER) task is introduced to help researchers assess the integrity of their papers and verify the clarity of each knowledge unit. A multi-task learning model utilizing Tucker decomposition and span-level attention mechanism is proposed to accurately identify terms and their integrity. Experimental results show the effectiveness of the model.
EXPERT SYSTEMS WITH APPLICATIONS
(2024)