Article
Anatomy & Morphology
Sakhr Ahmed Murshid
Summary: Lacunar-canalicular (LC) permeability and fluid flow-shear stress/drag force play important roles in mechanotransduction in bone tissue by inducing mechanical stimuli in osteocytes, modulating cellular functions, and determining bone adaptation. Alterations in LC structure may affect the ability of osteocytes to sense and translate mechanical signals, potentially contributing to bone remodelling. This review discusses recent studies on LC networks, their formation and transfer of mechanical stimuli, and changes in structure, functional permeability, and mechanotransduction resulting from age, pathology, and mechanical loading. Applications of vibration and low-intensity pulsed ultrasound in bone healthcare and regeneration fields are also presented.
Review
Endocrinology & Metabolism
Wen Sang, Ani Ural
Summary: The review summarizes recent findings on modifications in osteocyte lacunar and canalicular morphology under physiological and pathological conditions. Studies have confirmed a reduction in lacunar density with age and an increase in lacunar size with lactation. Changes in canalicular density, length, branching, and lacunar morphology have also been observed. Finite element models provide insights into how these modifications may affect fluid flow and local strains, ultimately influencing osteocyte mechanosensitivity.
CURRENT OSTEOPOROSIS REPORTS
(2023)
Article
Computer Science, Artificial Intelligence
Xiangde Luo, Guotai Wang, Tao Song, Jingyang Zhang, Michael Aertsen, Jan Deprest, Sebastien Ourselin, Tom Vercauteren, Shaoting Zhang
Summary: Segmentation of organs or lesions from medical images is crucial in clinical applications, and traditional methods as well as CNN have limitations. A novel deep learning-based interactive segmentation method is proposed to efficiently handle complex cases, reduce user interactions, and exhibit strong generalization capabilities.
MEDICAL IMAGE ANALYSIS
(2021)
Review
Endocrinology & Metabolism
G. Vahidi, C. Rux, V. D. Sherk, C. M. Heveran
Summary: Remodeling of the LCS can impact the morphology of the LCS itself, as well as the composition and mechanical properties of surrounding bone tissue. The effects of LCS remodeling on bone quality at physiologically-relevant length scales are still not fully understood due to limited studies evaluating tissue-scale bone properties near the LCS.
Article
Health Care Sciences & Services
Rodrigo Dalvit Carvalho da Silva, Thomas Richard Jenkyn, Victor Alexander Carranza
Summary: Segmentation is vital in medical imaging analysis for extracting regions of interest, and this study focuses on using a 3D CNN to segment skulls in MRI, creating standard volumetric labels from CT scans in STL models. The research finds that through training and manual corrections, the accuracy of skull segmentation can be significantly improved.
JOURNAL OF PERSONALIZED MEDICINE
(2021)
Article
Physics, Multidisciplinary
Mengnan Liu, Yu Han, Xiaoqi Xi, Linlin Zhu, Shuangzhan Yang, Siyu Tan, Jian Chen, Lei Li, Bin Yan
Summary: The resolution of 3D structure reconstructed by laboratory nanoCT is often affected by ambient temperature changes. Existing correction methods based on projection alignment are time-consuming and complex. In this study, a fast correction method named MD-Unet is proposed, which directly processes the reconstructed slices to improve correction performance. Experimental results show that MD-Unet significantly enhances the correction performance in nanoCT artifacts.
Article
Geography, Physical
Xinqiao Duan, Lin Li, Yong Ge, Bo Liu
Summary: The Voronoi diagram is a crucial geo-computing structure with various applications. This article proposes a new algorithm to compute the geodesic Voronoi diagram, breaking down the complex task into regular routines for accurate computation. Experimental results demonstrate the importance of the geodesic Voronoi diagram in geo-computing.
GISCIENCE & REMOTE SENSING
(2023)
Article
Endocrinology & Metabolism
Huiru Wang, Tianming Du, Rui Li, Russell P. Main, Haisheng Yang
Summary: This study investigated the interactive effects of different loading parameters on fluid velocity and shear stress within the osteocyte lacunar-canalicular system (LCS). The results showed that strain magnitude and rate were the main factors affecting velocity and shear stress, and their combination was not directly additive. The addition of a short rest between cycles enhanced the combined effect of these factors.
Review
Endocrinology & Metabolism
C. M. Heveran, J. D. Boerckel
Summary: Osteocytes directly modify the bone through resorption and deposition, a phenomena known as osteocyte osteolysis or LCS remodeling. Recent findings suggest that LCS remodeling is not only important for systemic calcium mobilization, but also for bone fracture resistance. Understanding the purpose and impact of LCS remodeling is crucial for therapeutic management of bone fragility.
CURRENT OSTEOPOROSIS REPORTS
(2023)
Article
Endocrinology & Metabolism
Hao Wang, Lilan Gao, Xuyi Chen, Chunqiu Zhang
Summary: This study investigated the mass transfer laws in bone microstructure under different gravity fields and found that high-intensity exercise and hypergravity can enhance the transport of solute molecules and nutrients in bone cells, while microgravity may inhibit mass transfer and lead to bone loss and osteoporosis.
JOURNAL OF BONE AND MINERAL METABOLISM
(2022)
Article
Computer Science, Artificial Intelligence
Feng Chen, Zhiheng Chen, Yuxiao Du, Zhuocheng Wu, Yuxing Li, Qi Hu
Summary: This study proposes a method of image deformation by body part size to reduce the cost and increase the personalized effect of virtual try-on. The method employs image segmentation algorithm to snap out the garment and adjusts its size and position according to the standard mannequin image. Experimental results demonstrate that this method can exhibit good performance in personalized try-on.
INTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS
(2022)
Article
Engineering, Biomedical
Emely Bortel, Liam M. Grover, Neil Eisenstein, Christian Seim, Heikki Suhonen, Alexandra Pacureanu, Peter Westenberger, Kay Raum, Max Langer, Francoise Peyrin, Owen Addison, Bernhard Hesse
Summary: Osteocyte lacunar-canalicular network (LCN) plays a crucial role in bone remodeling and mineral homeostasis, with a highly interconnected structure. The function of LCN is not only to optimize rapid access to bone mineral, but also to maintain high permeability when canaliculi interruptions occur. The structure of LCN can be influenced by anatomical location, subjected loads, and growth rate.
ADVANCED NANOBIOMED RESEARCH
(2022)
Article
Computer Science, Interdisciplinary Applications
Hao Wang, Jiaming Wang, Linwei Lyu, Shuping Wei, Chunqiu Zhang
Summary: This study established a multi-scale 3D osteon model and used the finite element method to numerically analyze mass transfer in the bone lacunar-canalicular system (LCS) under different gravity fields combined with high-intensity exercise. The results showed that hypergravity promoted mass transfer to deep lacunae, while microgravity inhibited it. High-intensity exercise increased the mass transfer rate in the LCS. This study provides a new strategy to combat and treat microgravity-induced osteoporosis.
COMPUTER METHODS IN BIOMECHANICS AND BIOMEDICAL ENGINEERING
(2023)
Article
Mathematics, Interdisciplinary Applications
Bing Zhao, Jiaqi Fan, Lifeng Xi
Summary: This study focuses on the geodesic metric of the limit space and the average distance limit of renormalized Lindstrom Snowflake networks, highlighting the differences from the Euclidean metric. By simplifying geodesic patterns using the dihedral group, a total of 10,850 patterns were obtained for calculation purposes.
FRACTALS-COMPLEX GEOMETRY PATTERNS AND SCALING IN NATURE AND SOCIETY
(2021)
Article
Computer Science, Artificial Intelligence
Pengfei Xian, Lai-Man Po, Jingjing Xiong, Chang Zhou, Yuzhi Zhao, Wing-Yin Yu, Weifeng Ou, Yujia Zhang, Xiaori Zhang
Summary: This study proposes a novel Pixel Voting Decoder to meet the performance requirements of both instance segmentation and semantic segmentation tasks. By regressing the interlayer pixel relationships between the input and output feature maps, the decoder dynamically decodes the higher level information from the encoder. The matrix computation for dynamic deconvolution enhances calculation efficiency.
EXPERT SYSTEMS WITH APPLICATIONS
(2022)
Article
Computer Science, Artificial Intelligence
Vien Ngoc Dang, Francesco Galati, Rosa Cortese, Giuseppe Di Giacomo, Viola Marconetto, Prateek Mathur, Karim Lekadir, Marco Lorenzi, Ferran Prados, Maria A. Zuluaga
Summary: Deep learning techniques for 3D brain vessel image segmentation have not been as successful as in other organ and tissue segmentation, mainly due to challenges in obtaining annotated training data and in segmenting relatively small vessels. To address these challenges, a novel annotation-efficient deep learning vessel segmentation framework has been proposed, which reduces the annotation burden by using weak patch-level labels.
MEDICAL IMAGE ANALYSIS
(2022)
Article
Multidisciplinary Sciences
Carles Bosch, Tobias Ackels, Alexandra Pacureanu, Yuxin Zhang, Christopher J. Peddie, Manuel Berning, Norman Rzepka, Marie-Christine Zdora, Isabell Whiteley, Malte Storm, Anne Bonnin, Christoph Rau, Troy Margrie, Lucy Collinson, Andreas T. Schaefer
Summary: The study integrates insights from physiology and structure, using a combination of different microscopy techniques to explore the function and structural features of biological tissues. The study identifies differences between two types of neurons, providing important clues for further research on neural circuits.
NATURE COMMUNICATIONS
(2022)
Review
Chemistry, Multidisciplinary
Francesco Galati, Sebastien Ourselin, Maria A. Zuluaga
Summary: Since the rise of deep learning in the mid-2010s, cardiac magnetic resonance image segmentation has made significant progress. However, the reliability and robustness of deep learning segmentation models still need attention. This paper studies the accuracy evolution of CMR segmentation, defines reliability and robustness, and identifies factors that limit the reliability and robustness of current deep learning CMR segmentation techniques. The paper also provides an overview of works focused on improving reliability and robustness, categorizing them into quality control methods and model improvement techniques.
APPLIED SCIENCES-BASEL
(2022)
Article
Radiology, Nuclear Medicine & Medical Imaging
Giacomo E. Barbone, Alberto Bravin, Alberto Mittone, Alexandra Pacureanu, Giada Mascio, Paola Di Pietro, Markus J. Kraiger, Marina Eckermann, Mariele Romano, Martin Hrabe de Angelis, Peter Cloetens, Valeria Bruno, Giuseppe Battaglia, Paola Coan
Summary: This study utilized multiscale X-PCI-CT technology to identify and quantify cellular and sub-cellular aging and neurodegeneration in deep neuronal and glial cell populations in a transgenic model of Alzheimer's disease. The study also demonstrated the localized and intracellular neuroprotective effects of pharmacological activation of mGlu2/3 receptors.
EUROPEAN JOURNAL OF NUCLEAR MEDICINE AND MOLECULAR IMAGING
(2022)
Article
Multidisciplinary Sciences
Lucas Pascal, Oscar J. Perdomo, Xavier Bost, Benoit Huet, Sebastian Otalora, Maria A. Zuluaga
Summary: This research presents a novel multi-task deep learning model for glaucoma diagnosis, leveraging the similarities of related eye-fundus tasks. The model simultaneously learns different segmentation and classification tasks, achieving higher performance with fewer parameters compared to training each task separately.
SCIENTIFIC REPORTS
(2022)
Review
Medicine, General & Internal
Fabien Lareyre, Hava Chaptoukaev, Sharon C. Kiang, Arindam Chaudhuri, Christian-Alexander Behrendt, Maria A. Zuluaga, Juliette Raffort
Summary: This review summarizes the applications of telemedicine in vascular surgery, highlighting the expected benefits, current limits, and future directions. Telemedicine can improve patient management through remote consultation, monitoring, and coaching, and has the potential to enhance access to care and contribute to the development of precision medicine.
JOURNAL OF CLINICAL MEDICINE
(2022)
Article
Computer Science, Artificial Intelligence
Yue Lu, Renjie Wu, Abdullah Mueen, Maria A. Zuluaga, Eamonn Keogh
Summary: Time series anomaly detection is a highly active research area in data mining, with many new approaches proposed each year. Despite the creative solutions, a twenty-year old technique called time series discord remains among the state-of-art techniques. Existing algorithms for computing time series discords have limitations in terms of online processing and scalability. In this work, a novel algorithm called DAMP is introduced to address these issues, allowing for the discovery of time series discords in datasets with trillions of data points.
DATA MINING AND KNOWLEDGE DISCOVERY
(2023)
Article
Computer Science, Artificial Intelligence
Julien Audibert, Pietro Michiardi, Frederic Guyard, Sebastien Marti, Maria A. Zuluaga
Summary: This study analyzes the performance of sixteen conventional, machine learning-based, and deep neural network approaches in detecting anomalies in multivariate time series, and finds that no family of methods outperforms the others in all cases.
PATTERN RECOGNITION
(2022)
Proceedings Paper
Acoustics
Madhu R. Kamble, Jose Patino, Maria A. Zuluaga, Massimiliano Todisco
Summary: The current outbreak of coronavirus is a serious global problem and has been declared a Public Health Emergency of International Concern. This study presents an automatic system for COVID-19 detection using breath, cough, and speech recordings. The results show promising accuracy using different auditory acoustic features.
2022 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP)
(2022)
Proceedings Paper
Computer Science, Theory & Methods
Matteo Guarrera, Baihong Jin, Tung-Wei Lin, Maria A. Zuluaga, Yuxin Chen, Alberto Sangiovanni-Vincentelli
Summary: We propose a simple yet effective method to improve the robustness of popular Out-of-Distribution (OoD) detection methods against label shift in deep neural networks. By considering class entry activation, we introduce a class-wise thresholding scheme that maintains similar OoD detection performance even in the presence of label shift in the test distribution.
2022 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION WORKSHOPS, CVPRW 2022
(2022)
Article
Multidisciplinary Sciences
Ruxandra Cojocaru, Oonagh Mannix, Marie Capron, C. Giles Miller, Pierre-Henri Jounea, Benoit Gallet, Denis Falconet, Alexandra Pacureanu, Stephen Stukins
Summary: The study reveals the structure and properties of the pollen wall, providing a deeper understanding of the evolutionary success of conifers and other plants over geological time.
Proceedings Paper
Audiology & Speech-Language Pathology
Madhu R. Kamble, Jose A. Gonzalez-Lopez, Teresa Grau, Juan M. Espin, Lorenzo Cascioli, Yiqing Huang, Alejandro Gomez-Alanis, Jose Patino, Roberto Font, Antonio M. Peinado, Angel M. Gomez, Nicholas Evans, Maria A. Zuluaga, Massimiliano Todisco
Summary: The COVID-19 pandemic has saturated global public health services, emphasizing the importance of early diagnosis for controlling virus spread and managing healthcare demand. The DiCOVA 2021 challenge organizers collected a database for COVID-19 diagnosis through coughing audio samples, with team PANACEA presenting an automatic detection system that achieved significant improvement over the baseline.
Proceedings Paper
Engineering, Biomedical
Laura M. Ferrari, Guy Abi Hanna, Paolo Volpe, Esma Ismailova, Francois Bremond, Maria A. Zuluaga
Summary: This study proposes to identify the optimal wearable EEG electrode set for EEG-based event detection and monitoring by training an autoencoder architecture in a one-class classification setup with different electrode combinations as input data. The model's performance was evaluated using the F-score, demonstrating that a wearable configuration with electrodes in the forehead and behind the ear was the chosen optimal set. The study shows the beneficial impact of a learning-based approach in designing wearable devices for real-life event-related monitoring.
2021 43RD ANNUAL INTERNATIONAL CONFERENCE OF THE IEEE ENGINEERING IN MEDICINE & BIOLOGY SOCIETY (EMBC)
(2021)
Proceedings Paper
Computer Science, Artificial Intelligence
Rosa Candela, Pietro Michiardi, Maurizio Filippone, Maria A. Zuluaga
Summary: Accurate travel products price forecasting is important for customers and companies, and a data-driven framework is introduced to continuously monitor and maintain deployed time-series forecasting models' performance. This helps guarantee stable performance of travel products price forecasting models.
MACHINE LEARNING AND KNOWLEDGE DISCOVERY IN DATABASES: APPLIED DATA SCIENCE TRACK, ECML PKDD 2020, PT IV
(2021)
Proceedings Paper
Computer Science, Artificial Intelligence
Lucas Pascal, Pietro Michiardi, Xavier Bost, Benoit Huet, Maria A. Zuluaga
Summary: Multi-task learning is popular due to resource usage and performance advantages. Joint optimization of parameters across multiple tasks remains an active research area. Maximum Roaming method, inspired by dropout, randomly varies parameter partitioning and forces visits to multiple tasks, providing superior regularization and performance impact.
THIRTY-FIFTH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE, THIRTY-THIRD CONFERENCE ON INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE AND THE ELEVENTH SYMPOSIUM ON EDUCATIONAL ADVANCES IN ARTIFICIAL INTELLIGENCE
(2021)