Article
Automation & Control Systems
Vignesh Sampath, Inaki Maurtua, Juan Jose Aguilar Martin, Andoni Rivera, Jorge Molina, Aitor Gutierrez
Summary: This article presents a method for improving the generalization ability of surface defect identification tasks by exploiting auxiliary information beyond the primary labels. By jointly learning features of pixel-level segmentation masks, object-level bounding boxes, and global image-level classification labels, the proposed method significantly improves the performance of state-of-the-art models. Experimental results show an overall accuracy of 97.1%, a Dice score of 0.926, and a mean average precision of 0.762 on defect classification, segmentation, and detection tasks.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2023)
Article
Chemistry, Analytical
Chung-Feng Jeffrey Kuo, Wei-Ren Wang, Jagadish Barman
Summary: This study presents a turnkey integrated system that can be operated in real time for textile manufacturers to recognize and classify eight types of defects in woven fabric. Four-stage image processing procedure is applied to detect and classify defects, with an overall defect recognition rate of 96.60% demonstrated in experimental results.
Article
Otorhinolaryngology
Jared A. Shenson, George S. Liu, Joyce Farrell, Nikolas H. Blevins
Summary: This study demonstrates the feasibility of using multispectral imaging and deep learning to enhance surgical vision for automated identification of normal human head and neck tissues. By developing and testing a novel multispectral imaging system and utilizing convolutional neural network machine learning models, the accuracy of tissue classification was significantly improved, outperforming traditional white-light imaging and surgical residents in tissue identification.
OTOLARYNGOLOGY-HEAD AND NECK SURGERY
(2021)
Article
Computer Science, Artificial Intelligence
Andreas Theissler, Mark Thomas, Michael Burch, Felix Gerschner
Summary: This paper proposes a model-agnostic approach ConfusionVis for evaluating and selecting multi-class classifiers based on their confusion matrices. By incorporating human knowledge and analyzing the per-class errors and class confusions, the proposed approach enables a more efficient training process and yields better models for specific applications.
KNOWLEDGE-BASED SYSTEMS
(2022)
Article
Computer Science, Information Systems
Ravail Singh, Varun Mumbarekar
Summary: The paper presents a new software based on artificial intelligence for automatically identifying benthic fauna through microscopic images, reducing the complexity of the identification process.
Article
Computer Science, Artificial Intelligence
Nuno Moniz, Vitor Cerqueira
Summary: In this paper, an Automated Imbalanced Classification method, ATOMIC, is proposed for imbalanced classification tasks. ATOMIC provides a ranking of solutions most likely to ensure an optimal approximation to a new domain, drastically reducing associated computational complexity and energy consumption by anticipating the loss of a large set of predictive solutions in new imbalanced learning tasks. Results demonstrate that the proposed method provides a relevant approach to imbalanced learning while reducing learning and testing efforts of candidate solutions by approximately 95%.
EXPERT SYSTEMS WITH APPLICATIONS
(2021)
Article
Green & Sustainable Science & Technology
Ke Zhang, Sarvesh Maskey, Hiromu Okazawa, Kiichiro Hayashi, Tamano Hayashi, Ayako Sekiyama, Sawahiko Shimada, Lameck Fiwa
Summary: This study compared three UAV image classification methods and found that each method has its own strengths and weaknesses, requiring selection based on specific needs. Combining different methods can monitor multiple resources or wastes simultaneously, contributing to an improved integrated resource management system.
Article
Environmental Sciences
Jan Conradt, Gregor Boerner, Angel Lopez-Urrutia, Christian Moellmann, Marta Moyano
Summary: The article introduces a processing pipeline called Dynamic Optimization Cycle (DOC), which systematizes and streamlines the model adaptation process in plankton image classification based on Deep Neural Networks (DNNs). By automatically updating the training dataset, the DOC pipeline enables strong performance and high throughput classification in plankton analysis at different locations or time periods.
FRONTIERS IN MARINE SCIENCE
(2022)
Article
Mathematics
Chaman Verma, Zoltan Illes, Deepak Kumar
Summary: Google Forms is a cutting-edge tool for gathering research data in the education field. This study proposes a novel predictive algorithm called Student Demographic Identification (SDI) to accurately identify students' demographic features. The proposed algorithm significantly improves the performance of traditional machine learning algorithms in predicting students' age-group, course, gender, locality, and institution.
Article
Computer Science, Software Engineering
Yaqin Zhou, Jing Kai Siow, Chenyu Wang, Shangqing Liu, Yang Liu
Summary: Security patches in open source software are crucial for protecting against cyber attacks, but the majority of vulnerabilities and their corresponding patches remain undisclosed. Curating and gathering these patches can be difficult due to their hidden nature.
ACM TRANSACTIONS ON SOFTWARE ENGINEERING AND METHODOLOGY
(2022)
Article
Construction & Building Technology
Peng Zhou, Yuan Chang
Summary: This study successfully automated the classification of building structures in urban environments using machine learning, with the Gradient Boosted Decision Tree algorithm showing the best performance. The classification method contributes to further research on the relationship between urban form, building structures, and resource requirements.
JOURNAL OF BUILDING ENGINEERING
(2021)
Article
Environmental Sciences
Nelle Meyers, Ana Catarino, Annelies M. Declercq, Aisling Brenan, Lisa Devriese, Michiel Vandegehuchte, Bavo De Witte, Colin Janssen, Gert Everaert
Summary: Microplastic pollution is a growing concern due to limited knowledge and difficulties in surveying and monitoring. An innovative approach combining high-throughput screening with automation has been developed to train machine learning models for plastic detection and polymer identification, demonstrating high accuracy in detecting and identifying plastic particles in environmental samples. This method represents a cost-effective, time-efficient, and reliable way to characterize microplastics.
SCIENCE OF THE TOTAL ENVIRONMENT
(2022)
Article
Engineering, Multidisciplinary
Ahmet Cagdas Seckin, Mine Seckin
Summary: A new feature extraction method for fabric defect detection is proposed, which is faster and more accurate compared to traditional texture feature extraction methods. This method can be used on low-level devices.
ALEXANDRIA ENGINEERING JOURNAL
(2022)
Article
Ecology
Rahim Azadnia, Kamran Kheiralipour
Summary: This study successfully classifies hawthorns according to their maturity levels using image processing and artificial intelligence techniques. By extracting features and applying machine learning algorithms, efficient classification of hawthorns is achieved.
ECOLOGICAL INFORMATICS
(2022)
Article
Dentistry, Oral Surgery & Medicine
J. L. Liu, S. H. Li, Y. M. Cai, D. P. Lan, Y. F. Lu, W. Liao, S. C. Ying, Z. H. Zhao
Summary: This study developed an automatic screening tool using deep learning to evaluate adenoid hypertrophy from lateral cephalograms. The results showed that the model had high accuracy and significantly improved speed and efficiency in assessing adenoid hypertrophy.
JOURNAL OF DENTAL RESEARCH
(2021)
Article
Computer Science, Software Engineering
Yuwei Zhang, Dahai Jin, Ying Xing, Yunzhan Gong
JOURNAL OF SYSTEMS AND SOFTWARE
(2020)
Article
Computer Science, Artificial Intelligence
Ying Xing, Xiaomeng Qian, Yu Guan, Bin Yang, Yuwei Zhang
Summary: Cross-project defect prediction is a popular research direction in software reliability, and traditional methods struggle in capturing the semantic and contextual information of programs. This paper applies technology from the NLP domain and proposes a deep learning model, Generative Adversarial Long-Short Term Memory Neural Networks (G-LSTM), to automatically learn the semantic and contextual features of programs. Experimental results show that the proposed method outperforms traditional and state-of-the-art methods in evaluation metrics.
PATTERN RECOGNITION LETTERS
(2022)
Article
Computer Science, Information Systems
Yanyang Zhao, Yawen Wang, Yuwei Zhang, Dalin Zhang, Yunzhan Gong, Dahai Jin
Summary: The ST-TLF framework proposed in this study can effectively perform cross-version defect prediction and improve the accuracy of CVDP. By selecting the best training set and eliminating concept drift, ST-TLF overcomes the limitations of previous research and shows significant improvements in various metrics.
INFORMATION AND SOFTWARE TECHNOLOGY
(2022)
Proceedings Paper
Computer Science, Software Engineering
Huiquan Gong, Yuwei Zhang
Summary: Researchers have utilized artificial intelligence to improve defect identification in software quality, proposing a feature selection method based on majority voting which achieved optimal performance at a ratio of 20%. This finding can serve as a practical guideline for software defect identification.
PROCEEDINGS OF 2021 IEEE 12TH INTERNATIONAL CONFERENCE ON SOFTWARE ENGINEERING AND SERVICE SCIENCE (ICSESS)
(2021)
Proceedings Paper
Computer Science, Artificial Intelligence
Huiquan Gong, Yuwei Zhang, Ying Xing, Wei Jia
PROCEEDINGS OF 2019 IEEE 10TH INTERNATIONAL CONFERENCE ON SOFTWARE ENGINEERING AND SERVICE SCIENCE (ICSESS 2019)
(2019)
Proceedings Paper
Computer Science, Software Engineering
Wei Jia, Yawen Wang, Yuwei Zhang, Yunzhan Gong
PROCEEDINGS OF 2018 IEEE 9TH INTERNATIONAL CONFERENCE ON SOFTWARE ENGINEERING AND SERVICE SCIENCE (ICSESS)
(2018)