Article
Engineering, Multidisciplinary
Julian N. Heidenreich, Colin Bonatti, Dirk Mohr
Summary: Mechanics-specific recurrent neural network (RNN) models can describe the complex three-dimensional stress-strain response of elasto-plastic solids for arbitrary loading paths. A strategy of training with datasets comprising random walks in strain space and transfer learning can significantly improve the generalization ability and convergence rates of the models. Ensemble transfer learning from multiple materials further enhances the accuracy and generalization ability of the models.
INTERNATIONAL JOURNAL FOR NUMERICAL METHODS IN ENGINEERING
(2023)
Article
Chemistry, Multidisciplinary
Yusuf Ozcevik
Summary: Smart environments require artificial intelligence for digitizing shopping processes, especially for the payment of unpackaged products. In this study, transfer learning is used to identify the type of unpackaged nuts, and the TL models achieve a promising validation accuracy of 96%.
APPLIED SCIENCES-BASEL
(2023)
Article
Computer Science, Information Systems
Sk Mahmudul Hassan, Arnab Kumar Maji, Michal Jasinski, Zbigniew Leonowicz, Elzbieta Jasinska
Summary: The study focuses on using deep convolutional neural network (CNN) models to identify and diagnose diseases in plants from their leaves. By achieving higher disease classification accuracy rates compared to traditional approaches, the implemented models show promise in efficient disease identification with less training time.
Article
Construction & Building Technology
Renner de Assis Garcia Sobrinho, Franklin Piauhy Neto, Henrique Fernandes
Summary: The research aimed to create a public database of images of cracks in mortar coating, considering different types of surface finish. The training accuracy varied based on surface finish and data balancing, with the scrapped type showing the lowest accuracy.
Article
Computer Science, Artificial Intelligence
Fatemeh Mohammadi Shakiba, Milad Shojaee, S. Mohsen Azizi, MengChu Zhou
Summary: Recent AI-based methods have shown promise in real-time detection and location estimation of transmission line faults. This paper introduces a novel transfer learning framework that utilizes a pre-trained convolutional neural network for diagnosing faults in transmission lines of different lengths and impedances.
Article
Computer Science, Information Systems
P. P. Fathimathul Rajeena, S. U. Aswathy, Mohamed A. Moustafa, Mona A. S. Ali
Summary: Farmers are not aware of the various corn diseases that affect agriculture, resulting in increasing crop failures due to lack of effective treatment or identification methods. Common corn diseases such as rust, blight, and northern leaf grey spot are prevalent. Accurate detection of diseases is not possible by visual observation alone, leading to improper pesticide use and potential harm to human health. Therefore, ensuring food security depends on accurate and automatic disease detection, which can be achieved by applying modern digital technologies.
Article
Computer Science, Information Systems
Mohd Saad Hamid, NurulFajar Abd Manap, Rostam Affendi Hamzah, Ahmad Fauzan Kadmin
Summary: This article surveys the algorithm frameworks related to stereo matching algorithm, dividing them into traditional and artificial intelligence (AI) frameworks. AI-based methods show higher accuracy compared to traditional methods, ranking high in standard benchmark datasets. Recent trends in solving computer vision problems lean towards using AI and machine learning tools, with a focus on deep learning frameworks related to convolutional neural networks (CNN).
JOURNAL OF KING SAUD UNIVERSITY-COMPUTER AND INFORMATION SCIENCES
(2022)
Review
Computer Science, Artificial Intelligence
Jia-Chi Wang, Yi-Chung Shu, Che-Yu Lin, Wei-Ting Wu, Lan-Rong Chen, Yu-Cheng Lo, Hsiao-Chi Chiu, Levent Ozcakar, Ke-Vin Chang
Summary: This study aimed to explore and summarize the performance of deep learning algorithms in the automatic sonographic assessment of the median nerve at the carpal tunnel level. The results showed that the deep learning algorithm enables automated localization and segmentation of the median nerve in ultrasound imaging with acceptable accuracy and precision. Future research should validate the performance of deep learning algorithms in detecting and segmenting the median nerve along its entire length and across datasets obtained from various ultrasound manufacturers.
ARTIFICIAL INTELLIGENCE IN MEDICINE
(2023)
Review
Biochemistry & Molecular Biology
Satish Vishwanathaiah, Hytham N. Fageeh, Sanjeev B. Khanagar, Prabhadevi C. Maganur
Summary: In the era of global epidemic, oral problems have a significant impact on a large population of children. Early diagnosis, prevention, and treatment of these disorders are crucial for children's optimal health. Artificial intelligence (AI) has made tremendous progress in recent years and infiltrated even traditionally human-specialist domains. AI models are frequently used in pediatric dentistry for accurate diagnosis, assisting clinicians and dentists in decision making, developing preventive strategies, and establishing treatment plans.
Review
Computer Science, Artificial Intelligence
Nripendra Kumar Singh, Khalid Raza
Summary: With the advancement of deep learning, significant progress has been made in dental and maxillofacial image analysis, enabling tasks such as dental structure segmentation, classification, and identification of common dental diseases with high accuracy. This progress holds promise for improved diagnosis and treatment planning in dentistry.
EXPERT SYSTEMS WITH APPLICATIONS
(2022)
Article
Biochemistry & Molecular Biology
Pang-Shuo Huang, Yu-Heng Tseng, Chin-Feng Tsai, Jien-Jiun Chen, Shao-Chi Yang, Fu-Chun Chiu, Zheng-Wei Chen, Juey-Jen Hwang, Eric Y. Chuang, Yi-Chih Wang, Chia-Ti Tsai
Summary: The use of artificial intelligence (AI) with electrocardiograms (ECGs) can identify significant coronary artery disease (CAD) and determine the site of the coronary obstruction. This technology can be easily implemented in health check-ups to identify high-risk patients for future coronary events.
Review
Chemistry, Analytical
Elizar Elizar, Mohd Asyraf Zulkifley, Rusdha Muharar, Mohd Hairi Mohd Zaman, Seri Mastura Mustaza
Summary: Most existing convolutional neural network-based deep-learning models suffer from spatial information loss and inadequate feature representation due to their inability to capture multiscale-context information and semantic information throughout pooling operations. Multiscale feature learning and fusion are crucial for optimal feature extraction and representation in deep-learning networks.
Article
Chemistry, Multidisciplinary
Arnaud Berenbaum, Herve Delingette, Aurelien Maire, Cecile Poret, Claire Hassen-Khodja, Stephane Breant, Christel Daniel, Patricia Martel, Lamiae Grimaldi, Marie Frank, Emmanuel Durand, Florent L. Besson
Summary: The study aimed to assess the feasibility of using a three-dimensional deep convolutional neural network (3D-CNN) for the general triage of whole-body FDG PET in daily clinical practice. A custom-made 3D-CNN was trained and tested on a well-balanced data sample, showing promising results. However, further studies are needed to overcome the challenges presented by real-life PET data.
APPLIED SCIENCES-BASEL
(2023)
Review
Cell Biology
Zuhui Zhang, Ying Wang, Hongzhen Zhang, Arzigul Samusak, Huimin Rao, Chun Xiao, Muhetaer Abula, Qixin Cao, Qi Dai
Summary: With the rapid development of computer technology, artificial intelligence (AI) has become increasingly prominent in ophthalmology research. Previous AI-related research in ophthalmology focused on screening and diagnosing fundus diseases, while the field has expanded to encompass ocular surface diseases. This review aims to summarize current AI research in ocular surface diseases such as pterygium, keratoconus, infectious keratitis, and dry eye to identify mature AI models and potential algorithms for future research.
FRONTIERS IN CELL AND DEVELOPMENTAL BIOLOGY
(2023)
Article
Chemistry, Multidisciplinary
Ritu Chauhan, Sahil Shighra, Hatim Madkhali, Linh Nguyen, Mukesh Prasad
Summary: Efficient waste management is crucial for reducing health and environmental risks. This study proposes a deep learning technique using a convolutional neural network to classify trash materials and improve waste management efficiency. The proposed method outperformed alternative methods in terms of performance.
APPLIED SCIENCES-BASEL
(2023)