Article
Construction & Building Technology
Tiejun Liu, Yutong Ju, Hanxiong Lyu, Qinglin Zhuo, Hanjie Qian, Ye Li
Summary: Due to the shortage and high price of river sand, the use of sea sand as a fine aggregate for concrete is being considered. Seashells are fragile and have an undesirable effect on the compressive strength of concrete. In this study, machine learning methods were used to analyze seashells and segment sea sand photos.
CASE STUDIES IN CONSTRUCTION MATERIALS
(2023)
Article
Oncology
Min-Jen Tsai, Ping-Yi Lin, Ming-En Lee
Summary: As we increasingly rely on advanced imaging for medical diagnosis, it is crucial that our computer programs accurately interpret these images. This study investigates how even small disruptions, such as changing a single pixel, can deceive our advanced algorithms. The findings highlight the concern about the reliability of current computer-aided diagnostic tools and the need for models that can resist such small disturbances.
Article
Geochemistry & Geophysics
Bona Hiu Yan Chow, Constantino Carlos Reyes-Aldasoro
Summary: This paper presents a computer-vision-based methodology for automatic image-based classification of gemstones, achieving higher accuracy and faster processing time compared to human experts. The study highlights the potential of computer vision in gemmology and related fields.
Article
Agronomy
Florian Kitzler, Helmut Wagentristl, Reinhard W. Neugschwandtner, Andreas Gronauer, Viktoria Motsch
Summary: Modern precision agriculture relies on stable computer vision outputs, where an important task is plant segmentation. The selection of training data is crucial for the quality of the resulting models, with plant cover and input features having a significant impact on segmentation quality. Decision tree classifiers with multi-feature input outperform threshold-based methods due to more balanced training data, and single-feature input decision tree classifiers can compete with state-of-the-art models when trained with the same data.
Article
Computer Science, Artificial Intelligence
Julian Luengo, Raul Moreno, Ivan Sevillano, David Charte, Adrian Pelaez-Vegas, Marta Fernandez-Moreno, Pablo Mesejo, Francisco Herrera
Summary: This paper reviews and categorizes computer vision techniques for metallographic image segmentation, introduces deep learning-based ensemble techniques utilizing pixel similarity, and conducts thorough comparisons in real-world datasets to discuss strengths, weaknesses, and application frameworks. The paper also addresses open challenges in the field to provide guidance for future research to fill existing gaps.
INFORMATION FUSION
(2022)
Article
Computer Science, Information Systems
Bernat Galmes, Gabriel Moya-Alcover, Pedro Bibiloni, Javier Varona, Antoni Jaume-i-Capo
Summary: This article presents a robust segmentation method for measuring toenails. The method is used in a clinical trial to objectively quantify the incidence of a specific pathology. It uses the Hough transform to locate the tip of the toe and estimate the nail location and size, and then classifies the super-pixels based on their geometric and photometric information. The watershed transform is then used to delineate the border of the nail.
MULTIMEDIA TOOLS AND APPLICATIONS
(2022)
Article
Biochemical Research Methods
Nikita Genze, Maximilian Wirth, Christian Schreiner, Raymond Ajekwe, Michael Grieb, Dominik G. Grimm
Summary: In this study, a combined deblurring and segmentation model called DeBlurWeedSeg is proposed for weed and crop segmentation in motion blurred images. The model outperforms a standard segmentation model without deblurring, showing a relative improvement of 13.4% in terms of the Sorensen-Dice coefficient.
Article
Computer Science, Artificial Intelligence
Daniela Medley, Carlos Santiago, Jacinto C. Nascimento
Summary: This paper presents a cyclic collaborative framework, CyCoSeg, which overcomes the limitation of deep neural networks in segmenting small objects in medical images. The framework combines a deep active shape model and a semantic segmentation network, and achieves competitive results in left ventricle segmentation as well as lung and kidney segmentation in CT scans.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2022)
Article
Radiology, Nuclear Medicine & Medical Imaging
Adam Berkley, Camillo Saueressig, Utkarsh Shukla, Imran Chowdhury, Anthony Munoz-Gauna, Olalekan Shehu, Ritambhara Singh, Reshma Munbodh
Summary: State-of-the-art deep learning models show promising performance on cross-institutional predictions. They considerably improve on previous models and can transfer knowledge to new types of brain tumors without additional modeling.
Article
Engineering, Chemical
Xue Wu, Yanmei Meng, Jinlai Zhang, Jing Wei, Xulei Zhai
Summary: The growth of cane sugar crystals is crucial for the production process. However, the overlap of sugar particles hampers the accurate identification of crystal shapes from images. To address this, we propose ASSugarNet, a deep neural network-based amodal segmentation method for cane sugar crystals. The network incorporates a local feature enhance module and a foreground boundary enhancement loss function to improve segmentation quality. It also uses an amodal segmentation module to predict contours and masks for occluded crystals. Experimental results demonstrate that ASSugarNet outperforms other networks in crystal segmentation, achieving an MIoU of 81.47% and an OA of 91.00%. The AP for adhesive particle segmentation is 32.25%, with an AP50 of 52.14%.
JOURNAL OF FOOD ENGINEERING
(2023)
Article
Surgery
Pieter De Backer, Jennifer A. Eckhoff, Jente Simoens, Dolores T. Mueller, Charlotte Allaeys, Heleen Creemers, Amelie Hallemeesch, Kenzo Mestdagh, Charles Van Praet, Charlotte Debbaut, Karel Decaestecker, Christiane J. Bruns, Ozanan Meireles, Alexandre Mottrie, Hans F. Fuchs
Summary: This study explores the requirements and methods of instrument annotation in surgical video and imaging data, and successfully implements a bottom-up approach for team annotation in two types of surgeries, laying the foundation for future AI projects in instrument detection, segmentation, and pose estimation.
SURGICAL ENDOSCOPY AND OTHER INTERVENTIONAL TECHNIQUES
(2022)
Article
Computer Science, Artificial Intelligence
Shuai Luo, Yujie Li, Pengxiang Gao, Yichuan Wang, Seiichi Serikawa
Summary: This paper reviews state-of-the-art image segmentation methods based on meta-learning, introducing the background and differences with other similar methods, discussing various types of meta-learning methods and their applications in image segmentation, conducting experimental comparisons, and highlighting future trends of meta-learning in image segmentation.
PATTERN RECOGNITION
(2022)
Article
Plant Sciences
Natalia Sapoukhina, Tristan Boureau, David Rousseau
Summary: Despite the lack of annotated datasets of plant images with disease lesions, this study proposes a novel methodology for generating fluorescent images of diseased plants with automated lesion annotation. The U-Net model trained purely by a synthetically generated dataset efficiently segments disease lesions on fluorescent images of plant leaves, achieving a recall of 0.793% and an average precision of 0.723% on an empirical fluorescent test dataset. The use of synthetic data can facilitate the application of deep learning methods in precision crop protection and improve the generalization ability of deep learning models.
FRONTIERS IN PLANT SCIENCE
(2022)
Article
Ecology
Frank Schindler, Volker Steinhage
Summary: Reliably detecting and tracking animals in wildlife videos is crucial for researchers to analyze animal behavior and recognize individuals. This paper introduces SWIFT, a novel multi-object tracking and segmentation pipeline that greatly improves the quality of instance masks and tracking accuracy compared to state-of-the-art methods, as validated on a self-created wildlife video dataset.
ECOLOGICAL INFORMATICS
(2022)
Article
Mathematics, Interdisciplinary Applications
Ayesha Sohail, Mohamed Abdelsabour Fahmy, Usama Ahmad Khan
Summary: Medical imaging visualizes the diseased part inside the patient's body using images, relying on various scientific and technological disciplines. This paper introduces a new hybrid machine learning approach to analyze medical images, presenting a novel algorithm for studying breast cancer images. The algorithm follows step-by-step stages to process, analyze, and classify the images. This research is significant for the field of particle physics imaging.
COMPUTATIONAL PARTICLE MECHANICS
(2022)
Article
Engineering, Industrial
Houssem Barkaoui, Helmi Ben Rejeb, Abdelwahed Barkaoui, Joao Manuel R. S. Tavares
Summary: This article presents DPMA, a multi-criteria decision-making process to assist engineering managers in making decisions during the development process of medical devices. The model involves weighting criteria, measuring device performance, and calculating a final score using the SAW method. A case study on the development of a femoral implant demonstrates the application of the DPMA process.
ENGINEERING MANAGEMENT JOURNAL
(2023)
Article
Computer Science, Information Systems
Vahid Hajihashemi, Abdoreza Alavi Gharahbagh, Azam Bastanfard, Hugo S. Oliveira, Goncalo Almeida, Zhen Ma, Joao Manuel R. S. Tavares
Summary: A hybrid method based on the AV1 codec and mathematical methods is proposed to improve compression performance and reconstruction quality. The proposed method shows better results in experiments and can also be used as a general compression method for 2D images.
MULTIMEDIA TOOLS AND APPLICATIONS
(2023)
Article
Mathematics, Interdisciplinary Applications
Zhenhua Yu, Ayesha Sohail, Maryam Jamil, O. A. Beg, Joao manuel R. S. Tavares
Summary: Fractal patterns, with their recursive and self-similar nature, provide a better understanding of natural patterns compared to Euclidean geometry. They have been extensively used in applied sciences, particularly in architecture and design, where computational methods for fractal generation serve as reliable tools for modeling and evaluation.
FRACTALS-COMPLEX GEOMETRY PATTERNS AND SCALING IN NATURE AND SOCIETY
(2023)
Editorial Material
Automation & Control Systems
Chinmay Chakraborty, Joao Manuel R. S. Tavares, Shaohua Wan, Houbing Herbert Song
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2023)
Article
Computer Science, Artificial Intelligence
Eleni Theodoridou, Luigi Cinque, Filippo Mignosi, Giuseppe Placidi, Matteo Polsinelli, Joao Manuel R. S. Tavares, Matteo Spezialetti
Summary: This article provides an overview of contemporary methods used for hand tracking and gesture recognition, including the use of contactless devices and multiple sensors to overcome limitations of monocular vision systems. It also introduces common steps, techniques, and algorithms used in developing these systems and predicts future trends.
IEEE TRANSACTIONS ON HUMAN-MACHINE SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Selim Reza, Marta Campos Ferreira, J. J. M. Machado, Joao Manuel R. S. Tavares
Summary: An autonomous vehicle can operate without human involvement and significantly reduce traffic congestion in an intelligent transportation system. The study proposes an improved SARSA model for managing autonomous vehicles by introducing a Gaussian function to regulate the weights updating mechanism effectively and suggesting the MaxAbs scaled state values instead of MinMax for efficient understanding of the traffic environment.
Article
Materials Science, Multidisciplinary
Paulo M. O. Silva, Mucio C. C. Filho, Jose A. da Cruz, Antonio J. M. Sales, Antonio S. B. Sombra, Joao Manuel R. S. Tavares
Summary: This article examines the effect of cold rolling and deformation on the pitting corrosion resistance of AISI 301LN and 316L stainless steels. The study combines various techniques to analyze the microstructures and conducts corrosion tests according to ASTM standards. The results show that the increase in cold rolling reduction leads to a higher content of martensite. The 316L steel demonstrates better pitting corrosion resistance due to its higher molybdenum content, while the 301LN steel is more susceptible to deformation-induced martensite formation.
Article
Materials Science, Multidisciplinary
Gabriel de Castro Coelho, Antonio Almeida Silva, Marco Antonio dos Santos, Jose J. M. Machado, Joao Manuel R. S. Tavares
Summary: In this study, the ductile fracture behavior of ASTM A516 Gr.70 pressure vessel steel was investigated, and the differences in estimating ASTM E1820 and ISO 12135 standards were assessed. The results showed that the mechanical properties and fracture toughness of the steel were dependent on the rolling direction, mainly attributed to perlite banding. Additionally, the differences in J(IC) determination were associated with the different blunting line slope estimations on each standard, highlighting the necessity of a work-hardening-based blunting line for each assessed material.
Review
Computer Science, Artificial Intelligence
Mariana Coelho, Martin Cerny, Joao Manuel R. S. Tavares
Summary: Alzheimer's disease is a progressive and irreversible neurodegenerative condition that affects memory, thinking, and behavior. Deep learning models show promise in diagnosing Alzheimer's using magnetic resonance images, but disease classification remains challenging. Developing more effective and innovative techniques is necessary.
Article
Materials Science, Multidisciplinary
Daniel J. Cruz, Rui L. Amaral, Abel D. Santos, Joao Manuel Tavares
Summary: Advanced high-strength steels (AHSS) are popular in the automotive industry due to their high strength, allowing for lighter car body structures. However, they have drawbacks such as edge cracking. The hole expansion test is commonly used to evaluate the formability of sheet metal materials, but accurately visualizing the first cracking is challenging. To address this, a digital image processing method is proposed to enhance the accuracy and efficiency of the hole expansion test, by detecting the appearance of the first edge cracks and providing insights into material behavior.
Article
Engineering, Biomedical
Helano M. B. F. Portela, Rodrigo de M. S. Veras, Luis H. S. Vogado, Daniel Leite, Paulo E. Ambrosio, Anselmo Cardoso de Paiva, Joao Manuel R. S. Tavares
Summary: Corneal ulcers, a common eye disease, can now be monitored more effectively using a segmentation method based on the U-Net Convolutional Neural Network architecture, resulting in promising results.
COMPUTER METHODS IN BIOMECHANICS AND BIOMEDICAL ENGINEERING-IMAGING AND VISUALIZATION
(2023)
Article
Engineering, Multidisciplinary
Boopathi Dhanasekaran, Jagatheesan Kaliannan, Anand Baskaran, Nilanjan Dey, Joao Manuel R. S. Tavares
Summary: This paper presents the performance of load frequency control (LFC) for isolated multiple sources of electric power-generating units using a proportional integral derivative (PID) controller. A thermal, hydro, and gas power-generating unit are integrated into the system. The PID controller is proposed as a subordinate controller to stabilize system performance during sudden demands on the power system. The particle swarm optimization (PSO) algorithm is used to obtain optimal gain values for the PID controller, with various cost functions used to optimize controller gain parameters. The results show that the PSO-PID controller delivers a faster response compared to conventional methods, with significant improvements over GA and DE-based PID controllers.
Article
Computer Science, Information Systems
Abdorreza Alavigharahbagh, Vahid Hajihashemi, Jose J. M. Machado, Joao Manuel R. S. Tavares, Vincenzo Moscato
Summary: This article proposes a hierarchical method for action recognition based on temporal and spatial features. It addresses challenges such as camera movement and sudden scene changes by using optical flow to detect and cancel camera movement. The method also reduces computational cost by limiting the search region for spatial processing. The proposed approach can improve the performance of current HAR systems as a preprocessing step.
Article
Chemistry, Multidisciplinary
Liliana Pinho, Andreia S. P. Sousa, Claudia Silva, Christine Cunha, Rubim Santos, Joao Manuel R. S. Tavares, Soraia Pereira, Ana Rita Pinheiro, Jose Felix, Francisco Pinho, Filipa Sousa, Augusta Silva
Summary: This study aims to analyze the coactivation of antagonist muscles in the thigh and ankle in post-stroke subjects during the sit-to-stand task. The results show that the coactivation of ankle joint muscles in post-stroke subjects is more dysfunctional, indicating that the distal segment may more accurately reflect central nervous system dysfunction in post-stroke subjects.
APPLIED SCIENCES-BASEL
(2023)