Article
Chemistry, Analytical
Caili Zhang, Yali Li, Longgang Xiang, Fengwei Jiao, Chenhao Wu, Siyu Li
Summary: This paper presents a novel intersection-first approach for generating road networks in old downtown areas based on low quality crowd-sourced vehicle trajectories. The method utilizes virtual representative points and the CFDP algorithm to improve intersection detection accuracy, and employs Delaunay triangulation network and adaptive link-fitting scheme to generate road networks, demonstrating remarkable performance in road network generation for old downtown areas.
Article
Chemistry, Multidisciplinary
Akmal Jahan Mohamed Abdul Cader, Jasmine Banks, Vinod Chandran
Summary: This paper introduces a solution for contact-based biometric applications that use finger and palm prints, proposing an algorithm that incorporates fingerprint features and Delaunay triangulation to overcome rotation and scaling effects. The algorithm demonstrates robustness in processing palm prints.
APPLIED SCIENCES-BASEL
(2023)
Article
Computer Science, Artificial Intelligence
Jinwon Lee, Sang-Uk Cheon, Jeongsam Yang
Summary: DenX-Conv is proposed for improving object classification accuracy while maintaining connectivity of points in raw point clouds. By extracting effective local geometric features and applying a densely connected network, a classification accuracy of 92.5% was achieved on the ModelNet40 dataset.
PATTERN RECOGNITION
(2021)
Article
Chemistry, Analytical
Ziyan Zhang, Yan Liu, Jiawei Tian, Shan Liu, Bo Yang, Longhai Xiang, Lirong Yin, Wenfeng Zheng
Summary: Feature-based 3D reconstruction and tracking technology is widely used in the medical field, enabling surgeons to achieve three-dimensional reconstruction and track soft tissue motion in minimally invasive surgery. This technology helps improve surgical accuracy and success rates.
Article
Clinical Neurology
E. I. S. Hofmeijer, C. O. Tan, F. van der Heijden, R. Gupta
Summary: Researchers tested ensemble learning for selecting the best artificial intelligence models for intracranial hemorrhage detection, but ensemble learning methods did not outperform the single best model.
AMERICAN JOURNAL OF NEURORADIOLOGY
(2023)
Article
Computer Science, Artificial Intelligence
David Novoa-Paradela, Oscar Fontenla-Romero, Bertha Guijarro-Berdinas
Summary: OCENCH is an intuitive, robust and efficient One-Class Classification algorithm that represents the target class using subdivisible and expandable non-convex hulls. It can handle non-convex and disjointed shapes, and the execution can be parallelized to reduce the execution time.
INFORMATION FUSION
(2023)
Article
Engineering, Electrical & Electronic
Zhongxing Zheng, Weiming Liu, Ruikang Liu, Liang Wang, Liang Mao, Qisheng Qiu, Guangzheng Ling
Summary: A sequential updatable anomaly detection (SUAD) framework is proposed, which utilizes the Robbin-Monro algorithm and a Mahalanobis distance calculation method based on principal component analysis. SUAD achieves competitive results on both self-built and public datasets, while reducing model size and memory usage.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY
(2022)
Article
Biochemistry & Molecular Biology
Elife Z. Bagci, Fatma Senguler-Ciftci, Unver Ciftci, Ayhan Demir
Summary: This study successfully differentiated different structural classes of proteins by combining the aspect ratio of each tetrahedron in alpha shapes with principal components analysis, achieving high accuracy in distinguishing R and T structures of hemoglobin. The method converts individual protein structures into points on a plane, showing promise for solving classification problems in machine learning.
PROTEINS-STRUCTURE FUNCTION AND BIOINFORMATICS
(2021)
Article
Environmental Sciences
Fubing Zhang, Qun Sun, Jingzhen Ma, Zheng Lyu, Zhekun Huang
Summary: This study proposes a polygonal building aggregation method that clusters visually conglutinated buildings and establishes the contours of the aggregated results through a minimum loop search. Experimental results show that this method effectively preserves the features of the aggregated results and improves their visual clarity.
GEOCARTO INTERNATIONAL
(2023)
Article
Computer Science, Information Systems
E. Rachdi, I. El Khadiri, Y. El merabet, Y. Rhazi, C. Meurie
Summary: This paper introduces a novel local feature extraction operator called MTSP, which is composed of two single-scale encoders, STP and SSP, designed based on a novel set theory pattern encoding scheme. Unlike other parametric texture operators, MTSP incorporates dynamic thresholds and can capture more detailed image information through the fusion of STP and SSP encoders. Experimental results demonstrate that MTSP achieves reliable performance stability on ten texture datasets and outperforms several representative methods in texture modeling, as verified by statistical tests.
JOURNAL OF KING SAUD UNIVERSITY-COMPUTER AND INFORMATION SCIENCES
(2023)
Article
Computer Science, Information Systems
Jingxue Wang, Xuetao Zheng, Zhenghui Xu
Summary: A quasi-dense matching algorithm with feature line constraint is proposed to address the deformation problem at the edges of artificial objects, achieving effective matching results.
Article
Computer Science, Artificial Intelligence
Wei Lin, Junyu Gao, Qi Wang, Xuelong Li
Summary: The study focuses on anomaly detection in various scenes, utilizing a synthetic anomaly event generating system to build a large dataset, detecting abnormal types using 3D CNN, and improving performance on real data through a cyclic 3D GAN for domain adaption.
Article
Computer Science, Theory & Methods
Junwei Zhou, Yijia Qian, Qingtian Zou, Peng Liu, Jianwen Xiang
Summary: This paper proposes a DeepSyslog method that represents Syslog with the context of log events and event metadata. It uses unsupervised sentence embedding to extract the semantic and context information hidden in the log stream, and combines it with event metadata to achieve high performance.
IEEE TRANSACTIONS ON INFORMATION FORENSICS AND SECURITY
(2022)
Article
Computer Science, Information Systems
Zirgham Ilyas, Zafar Aziz, Tehreem Qasim, Naeem Bhatti, Muhammad Faisal Hayat
Summary: The paper presents a hybrid deep network based approach for crowd anomaly detection in videos, using deep and handcrafted features, achieving high accuracy on the UMN crowd anomaly dataset and comparable performance on the challenging PETS 2009 dataset.
MULTIMEDIA TOOLS AND APPLICATIONS
(2021)
Article
Chemistry, Multidisciplinary
Antoine Badi Mame, Jules-Raymond Tapamo
Summary: Facial Expression Recognition (FER) is a rapidly growing field with diverse applications. This study analyzes the performance of various local descriptors and classifiers in the FER problem and identifies the best combinations. The results show that conventional FER approaches are still comparable to state-of-the-art deep learning methods.
APPLIED SCIENCES-BASEL
(2022)