Article
Multidisciplinary Sciences
Yuanyuan Qu, Xuesheng Li, Zhiliang Qin, Qidong Lu
Summary: In this paper, a novel approach based on a multi-branch three-dimensional (3D) convolution neural network (CNN) model is proposed for accurate acoustic scene classification (ASC). Multiple frequency-domain representations of signals are formed by utilizing expert knowledge on acoustics and discrete wavelet transformations (DWT). The proposed 3D CNN architecture, featuring residual connections and squeeze-and-excitation attentions (3D-SE-ResNet), effectively captures both long-term and short-term correlations in environmental sounds. Additionally, an auxiliary supervised branch based on the chromatogram of the original signal is incorporated to alleviate overfitting risks. Numerical evaluation on a large-scale dataset demonstrates the superior performance of the proposed multi-input multi-feature 3D-CNN architecture over state-of-the-art methods.
SCIENTIFIC REPORTS
(2022)
Article
Mathematics
Wei-Cheng Lin, Yi-Ren Yeh
Summary: This study explores extracting bit and byte-level sequences from malware executables and proposes an efficient one-dimensional CNN model for malware classification. Experimental results show that our proposed 1D CNN models outperform existing 2D CNN models for malware classification by providing better performance with smaller resizing bit/byte-level sequences and less computational cost.
Article
Astronomy & Astrophysics
Iulia Chifu, Ricardo Gafeira
Summary: The study discusses the impact of magnetic fields on various phenomena in the solar corona, proposing a 3D coronal loop reconstruction method based on machine learning.
ASTROPHYSICAL JOURNAL LETTERS
(2021)
Article
Geography, Physical
Xinyao Zhou, Wenzuo Zhou, Xiaoli Fu, Yichen Hu, Jinlian Liu
Summary: A new architecture called mobile 3D convolutional vision transformer (MDvT) is proposed to integrate 3D convolution with transformer models, achieving significant improvements in classification accuracy and model runtime. The MDvT reduces model parameters and accelerates model operation through inverted residual structure and square patch.
INTERNATIONAL JOURNAL OF DIGITAL EARTH
(2023)
Article
Biology
Ritesh Maurya, Vinay Kumar Pathak, Radim Burget, Malay Kishore Dutta
Summary: The study proposes a computer-aided bioimage classification method that fuses features extracted from various convolutional neural network architectures and selects discriminatory features through variance analysis and evolutionary feature selection. This approach achieves superior performance and high compression ratio, significantly reducing computational complexity in classifying bioimages.
COMPUTERS IN BIOLOGY AND MEDICINE
(2021)
Article
Environmental Sciences
Mohammed Q. Q. Alkhatib, Mina Al-Saad, Nour Aburaed, Saeed Almansoori, Jaime Zabalza, Stephen Marshall, Hussain Al-Ahmad
Summary: A novel method called Tri-CNN and a three-branch feature fusion approach are proposed to address the issue of insufficient training samples in hyperspectral image (HSI) classification. Experimental results demonstrate that the proposed method exhibits remarkable performance in terms of overall accuracy (OA), average accuracy (AA), and Kappa metrics when compared to existing methods.
Article
Multidisciplinary Sciences
Zhuangwen Wu, Zhiping Wan, Dongdong Ge, Ludan Pan
Summary: This study proposes a method for recognizing car engine sounds based on a deformable feature map residual network. It extracts the time-frequency image features using offset and convolutional layers, and fuses them with the Mel frequency cepstral coefficients. Experimental results show that the proposed method achieves a significantly higher accuracy under various operating conditions compared to existing methods.
SCIENTIFIC REPORTS
(2022)
Article
Computer Science, Interdisciplinary Applications
Ghazaleh Torabi, Amin Hammad, Nizar Bouguila
Summary: This paper investigates the advantages of a fully optimized method and an annotated untrimmed data set for activity recognition of construction workers. It proposes an improved version of YOWO to enhance detection performance and conducts a sensitivity analysis and a case study to compare different methods.
JOURNAL OF COMPUTING IN CIVIL ENGINEERING
(2022)
Article
Computer Science, Information Systems
Meng Lou, Runze Wang, Yunliang Qi, Wenwei Zhao, Chunbo Xu, Jie Meng, Xiangyu Deng, Yide Ma
Summary: In this study, a novel multi-level global-guided branch-attention network (MGBN) is proposed for breast masses classification, aiming to leverage the global contextual information to refine feature representation. Experimental results demonstrate that the proposed method achieves promising results on two public mammographic mass classification databases.
MULTIMEDIA TOOLS AND APPLICATIONS
(2021)
Article
Computer Science, Interdisciplinary Applications
He Zhang, Zhijing Shen, Zhenhang Lin, Liwei Quan, Liangfeng Sun
Summary: This study proposes an automatic hierarchical model for the classification and correlation of bridge surface images, which enhances the accuracy and efficiency of bridge inspection. The results demonstrate the effectiveness of the proposed method in multi-scale target classification.
COMPUTER-AIDED CIVIL AND INFRASTRUCTURE ENGINEERING
(2023)
Article
Environmental Sciences
Mark Little, Emma E. George, Milou G. I. Arts, Jade Shivak, Sean Benler, Joel Huckeba, Zachary A. Quinlan, Vittorio Boscaro, Benjamin Mueller, Ana Georgina Cobian Guemes, Maria Isabel Rojas, Brandie White, Daniel Petras, Cynthia B. Silveira, Andreas F. Haas, Linda Wegley Kelly, Mark J. A. Vermeij, Robert A. Quinn, Patrick J. Keeling, Pieter C. Dorrestein, Forest Rohwer, Ty N. F. Roach
Summary: The study explored microbial diversity, gene expression, and biochemistry in coral colonies, revealing no spatial patterns but distinct signatures of macroorganismal hosts in microbiome, transcriptome, and metabolome. Firmicutes were more abundant in coral microbiome, while certain bacterial RNA transcripts and metabolites were found ubiquitously. Machine learning analysis identified differences in microbial transcripts and metabolites between coral and competitor samples.
FRONTIERS IN MARINE SCIENCE
(2021)
Article
Mathematics, Interdisciplinary Applications
Adi Alhudhaif, Bandar Almaslukh, Ahmad O. Aseeri, Osman Guler, Kemal Polat
Summary: Skin cancer is a common cancer type that can have fatal implications. Early diagnosis is crucial for successful treatment. Developing computer-aided systems using deep learning networks shows promise for accurate detection and classification of skin lesions. This study proposes a deep learning approach with a soft attention mechanism and data balancing techniques to improve skin lesion classification performance.
CHAOS SOLITONS & FRACTALS
(2023)
Article
Multidisciplinary Sciences
Linfeng Sun, Zhongrui Wang, Jinbao Jiang, Yeji Kim, Bomin Joo, Shoujun Zheng, Seungyeon Lee, Woo Jong Yu, Bai-Sun Kong, Heejun Yang
Summary: This study presents a sensor reservoir computing approach for language learning, utilizing two-dimensional memristors to achieve high dimensionality, nonlinearity, and fading memory. With an accuracy of 91% in classifying short sentences, it offers a low training cost and real-time solution for processing temporal and sequential signals in machine learning applications at the edge.
Article
Computer Science, Artificial Intelligence
Raluca Jalaboi, Frederik Faye, Mauricio Orbes-Arteaga, Dan Jorgensen, Ole Winther, Alfiia Galimzianova
Summary: Dermatological diagnosis automation is crucial for addressing the high prevalence of skin diseases and shortage of dermatologists. DermX and DermX+ are two explainable automated dermatological diagnosis methods that achieve near-expert diagnosis performance while providing expert-level explanations.
MEDICAL IMAGE ANALYSIS
(2023)
Article
Multidisciplinary Sciences
Sajib Saha, Janardhan Vignarajan, Shaun Frost
Summary: This paper presents a computationally efficient and memory efficient CNN-based system for automated detection of glaucoma. The system achieves high accuracy and speed while minimizing resource requirements. It performs well in classifying glaucomatous and non-glaucomatous images, making it suitable for integration into portable fundus cameras.
SCIENTIFIC REPORTS
(2023)
Article
Limnology
Matthew H. Long, Kevin Sutherland, Scott D. Wankel, David J. Burdige, Richard C. Zimmerman
LIMNOLOGY AND OCEANOGRAPHY
(2020)
Article
Plant Sciences
Sonia Blanco-Ameijeiras, Heather M. Stoll, Hongrui Zhang, Brian M. Hopkinson
JOURNAL OF PHYCOLOGY
(2020)
Article
Fisheries
Steve S. Doo, Andrea Kealoha, Andreas Andersson, Anne L. Cohen, Tacey L. Hicks, Zackary Johnson, Matthew H. Long, Paul McElhany, Nathaniel Mollica, Kathryn E. F. Shamberger, Nyssa J. Silbiger, Yuichiro Takeshita, D. Shallin Busch
ICES JOURNAL OF MARINE SCIENCE
(2020)
Article
Limnology
Daniel P. Owen, Matthew H. Long, William K. Fitt, Brian M. Hopkinson
Summary: This study estimated the contribution of different primary producers to the overall primary production on coral reefs, finding that the main producers varied between degraded and intact reef sites. By using a bottom-up approach, the researchers were able to provide more accurate estimates of production rates and validate the method's reliability through comparison with in situ measurements.
LIMNOLOGY AND OCEANOGRAPHY
(2021)
Article
Geochemistry & Geophysics
Hongrui Zhang, Sonia Blanco-Ameijeiras, Brian M. Hopkinson, Stefano M. Bernasconi, Luz Maria Mejia, Chuanlian Liu, Heather Stoll
Summary: In this study, a new method was explored to detect carbonic anhydrase (CA) activity within coccolithophores, revealing the importance of CA in the calcification pathway. It was found that there is no significant difference in the CA activity between a high and low CO2 treatment for E. huxleyi, but under low CO2 treatment the bicarbonate pumping rate was enhanced.
GEOCHIMICA ET COSMOCHIMICA ACTA
(2021)
Article
Oceanography
Matthew H. Long
Summary: The study shows that waves in shallow waters can introduce biases in aquatic biogeochemical EC measurements, while larger measurement heights can create a spectral gap between turbulence and wave frequencies, reducing wave-bias. The new analysis framework and measurement guidelines can effectively remove wave-bias, but wave-bias still exists in traditional EC analysis results.
JOURNAL OF GEOPHYSICAL RESEARCH-OCEANS
(2021)
Article
Green & Sustainable Science & Technology
Dirk Koopmans, Volker Meyer, Allison Schaap, Marius Dewar, Paul Farber, Matthew Long, Jonas Gros, Douglas Connelly, Moritz Holtappels
Summary: The study used pH eddy covariance to detect and quantify controlled release of CO2 on the seafloor, finding that the emission was significantly higher than proton flux from natural mineralization, and highlighting the importance of considering carbonate system kinetics in emission estimation.
INTERNATIONAL JOURNAL OF GREENHOUSE GAS CONTROL
(2021)
Article
Limnology
Jeff Coogan, Jennie E. Rheuban, Matthew H. Long
Summary: This study compares the measurements of the gradient flux (GF) method with two other methods, eddy covariance and benthic chambers, to show their strengths, weaknesses, and uncertainties. The results highlight the importance of sufficient DO gradient, consistent methods for comparison, and careful estimation of sensor scale and placement in complex bottom types when using the GF method.
LIMNOLOGY AND OCEANOGRAPHY-METHODS
(2022)
Article
Environmental Sciences
Matthew H. H. Long, Jordan W. Mora
Summary: Coastal nutrient pollution, or eutrophication, often occurs due to human activities in terrestrial watersheds, leading to degraded water quality over time. The management and monitoring of estuarine systems usually lag behind environmental degradation. In the case of Waquoit Bay National Estuarine Research Reserve, the loss of eelgrass meadows and the decline in macroalgal biomass have resulted in shifts in water quality and ecosystem structure, with an increase in phytoplankton biomass and a shift towards pelagic dominance. This shift may have wider implications for other eutrophic and warming estuaries in the future.
ESTUARIES AND COASTS
(2023)
Article
Limnology
Solomon T. Chen, Collin P. Ward, Matthew H. Long
Summary: Pelagic photosynthesis and respiration play crucial roles in controlling dissolved oxygen concentration in seawater. To address the lack of data on marine primary production, a novel automated water incubation system was developed to measure in situ rates of photosynthesis and respiration, providing high-resolution data on the metabolic state of pelagic ecosystems.
LIMNOLOGY AND OCEANOGRAPHY-METHODS
(2023)
Article
Limnology
Jeff Coogan, Matthew H. H. Long
Summary: The aquatic eddy covariance (AEC) technique is a versatile tool for understanding benthic fluxes. This paper evaluates the design and deployment of a long-term eddy covariance system (LECS) that provided reliable data for 6 months in a temperate seagrass meadow. The study found a gradual reduction in sensor response time, likely due to fouling, and introduced new spectral analysis techniques for real-time monitoring.
LIMNOLOGY AND OCEANOGRAPHY-METHODS
(2023)
Article
Multidisciplinary Sciences
Kewei Yu, Peiwei Liu, Dipna Venkatachalam, Brian M. Hopkinson, Karl F. Lechtreck