Article
Engineering, Biomedical
Zeyu Chen, Senyang Chen, Fengjun Hu
Summary: Due to the limitations in current deep learning models for segmenting dental CBCT images, this study proposed a CNN-Transformer Architecture UNet network that combines CNN and Transformer to effectively extract local features and capture long-range dependencies. Multiple spatial attention modules were included to enhance spatial information extraction. A novel Masked image modeling method was introduced to pre-train the CNN and Transformer modules simultaneously, mitigating limitations caused by insufficient labeled training data. Experimental results demonstrated superior performance in dental CBCT image segmentation, with real-world applicability in orthodontics and dental implants.
PHYSICS IN MEDICINE AND BIOLOGY
(2023)
Article
Computer Science, Artificial Intelligence
Ibrahim Delibasoglu
Summary: Building segmentation from aerial images is crucial for urban planning and population estimation. UNET, combined with Inception blocks, implements a two-level Inception approach to enhance feature extraction and prevent overfitting. The proposed INCSA-UNET architecture shows significant improvement in terms of F1 and Kappa quantitative measures without significantly increasing the number of parameters.
COMPUTING AND INFORMATICS
(2021)
Article
Environmental Sciences
Masoomeh Gomroki, Mahdi Hasanlou, Peter Reinartz
Summary: Change detection in urban areas is important for urban resource management and smart city planning. In this research, a semi-transfer learning method called EffV2 T-Unet was proposed for binary change detection in urban environments, using EfficientNetV2 T as the encoder and Unet as the decoder. The method was evaluated using two datasets and achieved an overall accuracy of 97.66%. The results demonstrated its effectiveness compared to other methods.
Article
Agriculture, Multidisciplinary
Amirhossein Zaji, Zheng Liu, Gaozhi Xiao, Pankaj Bhowmik, Jatinder S. Sangha, Yuefeng Ruan
Summary: The study aims to develop a state-of-the-art framework for localizing and counting wheat spikes using dotted annotation datasets in conjunction with dense map generation algorithms. The performance of wheat counting was improved by employing a hybrid density map estimation-non-maximal supervision algorithm.
COMPUTERS AND ELECTRONICS IN AGRICULTURE
(2022)
Article
Computer Science, Information Systems
Ekam Singh Chahal, Aarya Patel, Ayush Gupta, Archana Purwar, G. Dhanalekshmi
Summary: Prostate cancer is one of the most prevalent types of tumors in males worldwide, with age and family history being the main risk factors. MRI is highly recommended for detecting and localizing prostate cancer. This paper presents an automatic segmentation model for prostate regions in MRI scans based on Unet and Xception net, with the addition of local residual connections in the decoder stage of the Unet to enhance model performance. Experimental results demonstrate that the proposed model outperforms other methods studied.
MULTIMEDIA TOOLS AND APPLICATIONS
(2022)
Article
Engineering, Electrical & Electronic
Qiaoning Yang, Xiaodong Ji
Summary: This paper proposes an automatic pixel-level crack detection method based on deep transfer learning, which includes crack recognition and semantic segmentation stages, achieving efficient and reliable detection results. Experimental results show that the method can effectively detect pixel-level cracks and perform well in large-scale crack detection tasks.
IEEE SENSORS JOURNAL
(2021)
Article
Computer Science, Artificial Intelligence
Pradeep Kumar Roy, Fenish Umeshbhai Mali
Summary: This research aims to develop a model to help prevent image-based cyberbullying issues on social platforms. The experimental results show that the transfer learning-based model can accurately detect most cyberbullying posts.
COMPLEX & INTELLIGENT SYSTEMS
(2022)
Article
Computer Science, Artificial Intelligence
Fares Bougourzi, Cosimo Distante, Fadi Dornaika, Abdelmalik Taleb-Ahmed
Summary: Medical imaging has been widely used to analyze Covid-19 since its emergence in late 2019. This paper proposes the PAtt-Unet and DAtt-Unet architectures to improve the performance of Covid-19 infection segmentation. Experimental results show that both PAtt-Unet and DAtt-Unet enhance the performance of Att-Unet in segmenting Covid-19 infections.
MEDICAL IMAGE ANALYSIS
(2023)
Article
Geochemistry & Geophysics
Weibin Song, Xuping Feng, Gaoxiong Wu, Gongheng Zhang, Ying Liu, Xiaofei Chen
Summary: The neural network Res-Unet++ can automatically and accurately extract fundamental dispersion curves and overtones from F-J dispersion spectra after training, showing high accuracies in synthetic and real data. The network's effectiveness in extracting dispersion curves and adaptability through transfer learning have been demonstrated, providing advantages in generating more effective dispersion points.
JOURNAL OF GEOPHYSICAL RESEARCH-SOLID EARTH
(2021)
Article
Automation & Control Systems
Sadia Showkat, Shaima Qureshi
Summary: Due to the effect of COVID-19 on pulmonary tissues, Chest X-ray (CXR) and Computed Tomography (CT) images have become the preferred imaging methods for early detection of COVID-19 infections. The use of Convolutional Neural Networks (CNN) and Transfer Learning (TL) approach, specifically the ResNet architecture, has shown effectiveness in accurately classifying pneumonia cases from CXR images. The customized ResNet model achieved high global accuracy, precision, specificity, and sensitivity, providing reliable analysis of CXR images for clinical decision-making.
CHEMOMETRICS AND INTELLIGENT LABORATORY SYSTEMS
(2022)
Article
Computer Science, Artificial Intelligence
Rayyan Azam Khan, Yigang Luo, Fang-Xiang Wu
Summary: In this study, a deep-learning-based model is proposed for precise segmentation in hepatocellular carcinoma and metastasis clinical diagnosis. The model utilizes multi-scale approach, novel objective function, and extensive validation to achieve competitive performance. The results demonstrate the applicability and effectiveness of the proposed methodology in relevant medical segmentation applications.
ARTIFICIAL INTELLIGENCE IN MEDICINE
(2022)
Article
Mathematics
Rukesh Prajapati, Goo-Rak Kwon
Summary: Proper analysis of changes in brain structure is crucial for diagnosing brain disorders accurately. In this study, a novel architecture combining three parallel UNets and a residual network is proposed to improve upon baseline methods. By using three consecutive images as input, compressing and decompressing them individually, and enhancing image features with skip connections, the proposed architecture outperforms single conventional UNet and other UNet variants in segmentation accuracy.
Article
Mathematics
Hui Zhu, Shi Shu, Jianping Zhang
Summary: This paper proposes a novel variational-model-informed network (FAS-UNet) that uses model and algorithm priors to extract multiscale features, and it achieves strong competitiveness with other state-of-the-art methods in medical image segmentation tasks.
Article
Agronomy
J. Andrew, Jennifer Eunice, Daniela Elena Popescu, M. Kalpana Chowdary, Jude Hemanth
Summary: In this study, pre-trained models based on convolutional neural networks were used for efficient plant disease identification. The hyperparameters of popular pre-trained models were fine-tuned, and experiments were conducted using the PlantVillage dataset. The results demonstrated that DenseNet-121 achieved superior classification accuracy.
Article
Computer Science, Interdisciplinary Applications
Hsiu-Hsia Lin, Wen-Chung Chiang, Chao-Tung Yang, Chun-Tse Cheng, Tianyi Zhang, Lun-Jou Lo
Summary: This study introduced a convolutional neural network with a transfer learning approach for facial symmetry assessment based on 3-dimensional features to assist physicians in enhancing medical treatments. By transforming 3D features and applying data augmentation methods, a new model for evaluating facial symmetry was successfully trained, achieving satisfactory experimental results.
COMPUTER METHODS AND PROGRAMS IN BIOMEDICINE
(2021)
Article
Environmental Sciences
Bridget R. Scanlon, Ashraf Rateb, Assaf Anyamba, Seifu Kebede, Alan M. MacDonald, Mohammad Shamsudduha, Jennifer Small, Alexander Sun, Richard G. Taylor, Hua Xie
Summary: Water resources management in Africa is critical. This study assesses the spatiotemporal variability in water storage and its controls in major African aquifers. The results show declining trends in water storage in northern Africa due to irrigation water use, while rising trends are found in western Africa due to land use change and increased recharge. Climate extremes strongly control water storage in eastern and southern Africa.
ENVIRONMENTAL RESEARCH LETTERS
(2022)
Article
Environmental Sciences
Zhen Li, Zizhan Zhang, Bridget R. Scanlon, Alexander Y. Sun, Yun Pan, Shuqing Qiao, Hansheng Wang, Qiuyang Jia
Summary: This study developed an improved method to estimate sediment input changes in the Bohai Sea using GRACE data and satellite altimetry. The results revealed seasonal variations in sediment input and the contribution of coastal erosion.
SCIENCE OF THE TOTAL ENVIRONMENT
(2022)
Article
Environmental Sciences
Wenting Yang, Di Long, Bridget R. Scanlon, Peter Burek, Caijin Zhang, Zhongying Han, James J. Butler, Yun Pan, Xiaohui Lei, Yoshihide Wada
Summary: The North China Plain has experienced groundwater overexploitation due to rapid socio-economic development and irrigation demand. The operation of the South-to-North Water Diversion Project has provided an opportunity to sustain groundwater resources. This study used a high-resolution model to simulate and project groundwater storage in the region, and found that water diversion and reductions in water use could increase groundwater storage.
WATER RESOURCES RESEARCH
(2022)
Article
Geochemistry & Geophysics
Harpreet Kaur, Zhi Zhong, Alexander Sun, Sergey Fomel
Summary: Geologic carbon sequestration involves injecting captured carbon dioxide into subsurface formations for long-term storage. This study presents a deep-learning framework for monitoring CO2 saturation and determining the geologic controls on storage. The trained model accurately estimates CO2 saturation values and plume extent using time-lapse seismic data.
INTERPRETATION-A JOURNAL OF SUBSURFACE CHARACTERIZATION
(2022)
Article
Environmental Sciences
Jingyu Kang, Yang Lu, Yan Li, Zizhan Zhang, Hongling Shi
Summary: This study proposes an iteration method to investigate the variation of Antarctic basal water storage based on satellite observations and models. The research finds that basal water storage increased during 2003-2009, mainly in regions with active subglacial lakes. However, there are uncertainties in the study.
Article
Energy & Fuels
Svetlana A. Ikonnikova, Bridget R. Scanlon, Sofia A. Berdysheva
Summary: This paper fills the gap in existing research by examining the prerequisites for hydrogen trade in the context of global carbon neutrality and the impact of regional energy characteristics and regulations on the hydrogen market. It provides a comprehensive energy system perspective at a global scale.
Article
Environmental Sciences
Poulomee Coomar, Kousik Das, Palash Debnath, Swati Verma, Prerona Das, Ashis Biswas, Abhijit Mukherjee
Summary: This study investigates the role of submarine groundwater discharge in transporting arsenic from the Ganges river delta to the Bay of Bengal. The results reveal the presence of a plume carrying up to 30 μg/L of dissolved arsenic towards the sea. Arsenic distribution and transport are controlled by the Fe-Mn redox cycle and influenced by terrestrial groundwater discharge.
ENVIRONMENTAL POLLUTION
(2023)
Article
Environmental Sciences
Soumyajit Sarkar, Abhijit Mukherjee, Madhumita Chakraborty, Md Tahseen Quamar, Srimanti Duttagupta, Animesh Bhattacharya
Summary: Elevated fluoride in groundwater is a serious problem in India, affecting the health of a large population that relies on groundwater. This study explores the relationship between tectonics and fluoride distribution in groundwater using machine learning models. The random forest model is found to be the most accurate, and two high-risk areas are identified.
ENVIRONMENTAL SCIENCE AND POLLUTION RESEARCH
(2023)
Editorial Material
Engineering, Environmental
Abhijit Mukherjee, Sachchida Nand Tripathi, Kirpa Ram, Dipankar Saha
ENVIRONMENTAL SCIENCE & TECHNOLOGY LETTERS
(2023)
Article
Engineering, Civil
Soumendra N. Bhanja, Ethan T. Coon, Dan Lu, Scott L. Painter
Summary: Fully distributed, integrated surface-subsurface hydrological models (ISSHMs) have gained attention due to advancements in software, computing facilities, and data products. This study evaluates the performance of the Advanced Terrestrial Simulator (ATS) across seven diverse catchments in the US without calibration. Good performance is observed for streamflow and evapotranspiration in most catchments, with improvements seen when incorporating local sub-surface properties. ATS performs comparably to a calibrated semi-distributed model for streamflow and better for evapotranspiration. This study boosts confidence in the ISSHM models and community data products for understanding watershed function.
JOURNAL OF HYDROLOGY
(2023)
Article
Environmental Sciences
Shuo Zheng, Zizhan Zhang, Haoming Yan, Yaxian Zhao, Zhen Li
Summary: Our study reveals that the extreme change of water storage in the Yangtze River Basin has a significant impact on identifying the characteristics of drought events. We propose a new framework to reconstruct the pre-2003 total water storage anomaly using the nonlinear autoregressive with exogenous input (NARX) model. Results show that the frequency, duration, and average recovery time of drought events have significantly increased since 2000 in the Yangtze River Basin.
SCIENCE OF THE TOTAL ENVIRONMENT
(2023)
Editorial Material
Environmental Sciences
Abhijit Mukherjee, Soumendra N. Bhanja, Matthew Rodell, Yoshihide Wada, Prangaditya Malakar, Dipankar Saha, Alan M. MacDonald
Article
Geosciences, Multidisciplinary
Alexander Y. Sun, Peishi Jiang, Zong-Liang Yang, Yangxinyu Xie, Xingyuan Chen
Summary: Rivers and river habitats worldwide are facing sustained pressure from human activities and global environmental changes. It is crucial to quantify and manage river states in a timely manner for public safety and natural resource protection. This study presents a multistage, physics-guided, graph neural network (GNN) approach for basin-scale river network learning and streamflow forecasting.
HYDROLOGY AND EARTH SYSTEM SCIENCES
(2022)