Article
Construction & Building Technology
Yiqing Liu, Justin K. W. Yeoh
Summary: This paper proposes a robust crack segmentation approach using image patches, which combines an active contour model, convolutional neural network, and morphological operations to accurately detect and segment cracks. Experimental validation shows significant improvement in accuracy and robustness compared to previous work, with lower data labeling requirements.
AUTOMATION IN CONSTRUCTION
(2021)
Article
Computer Science, Artificial Intelligence
Mingbao Lin, Rongrong Ji, Xiaoshuai Sun, Baochang Zhang, Feiyue Huang, Yonghong Tian, Dacheng Tao
Summary: In this paper, a novel supervised online hashing scheme called FCOH is proposed to address the adaptivity and efficiency issues in online image hashing. By introducing a novel and efficient inner product operation, FCOH achieves fast online adaptivity and efficiency through class-wise updating and semi-relaxation optimization.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2022)
Article
Computer Science, Interdisciplinary Applications
Shahrizan Jamaludin, Ahmad Faisal Mohamad Ayob, Mohd Faizal Ali Akhbar, Ahmad Ali Imran Mohd Ali, Md Mahadi Hasan Imran, Syamimi Mohd Norzeli, Saiful Bahri Mohamed
Summary: Iris recognition is a powerful biometric system that is user-friendly, accurate, fast, and reliable, making it suitable for use during the COVID-19 pandemic. However, it still faces challenges such as pupil deformation, motion blur, eyelids blocking, reflection occlusion, and eyelashes obscure. Accurate segmentation of the pupillary boundary is crucial for successful iris recognition.
ADVANCES IN ENGINEERING SOFTWARE
(2023)
Article
Chemistry, Multidisciplinary
Liduo Liu, Yongji Long, Guoning Li, Ting Nie, Chengcheng Zhang, Bin He
Summary: In this study, a fast and accurate visual object tracking method based on Siamese networks is proposed, which incorporates multi-layer feature information and pixel-level correlation operation to enhance the feature extraction capability of the network. The algorithm improves the precision and success rates on various datasets and performs better in complex scenes such as occlusion, illumination changes, and fast-motion situations.
APPLIED SCIENCES-BASEL
(2023)
Article
Computer Science, Information Systems
Hasan Salman, Amir Hossein Taherinia, Davood Zabihzadeh
Summary: Content retrieval systems aim to retrieve images similar to a query image from a large dataset by using feature extraction and similarity measures. This paper presents a method called Deep Muti-teacher Transfer Hash (DMTH) that uses knowledge from multiple complex models to teach a simple one, utilizing an attention mechanism to obtain richer features and improving image retrieval performance without increasing evaluation time. Experimental results show that DMTH outperforms state-of-the-art methods, achieving higher accuracy and boosting the performance of student models.
MULTIMEDIA TOOLS AND APPLICATIONS
(2023)
Article
Computer Science, Information Systems
Aman Dureja, Payal Pahwa
Summary: With the continuous growth of medical image repositories, there is a need for an effective image retrieval system to simplify the process. Traditional image retrieval frameworks suffer from the poor extraction of low and high-level features, which creates a semantic gap. This research proposes a modified Convolutional Neural Network (CNN) approach for the retrieval of medical images, achieving accurate retrieval with an accuracy rate of around 94%.
MULTIMEDIA TOOLS AND APPLICATIONS
(2023)
Article
Cell Biology
Yutong Han, Zhan Zhang, Yafeng Li, Guoqing Fan, Mengfei Liang, Zhijie Liu, Shuo Nie, Kefu Ning, Qingming Luo, Jing Yuan
Summary: Automated evaluation of all glomeruli throughout the whole kidney is crucial for studying kidney function and understanding kidney disease mechanisms. This study proposes a deep learning-based segmentation method called FastCellpose to efficiently segment all glomeruli in whole mouse kidneys. The method shows superior performance compared to other cellular segmentation methods and significantly improves processing speed. By using this high-performance framework, the researchers were able to quantitatively analyze the developmental changes of mouse glomeruli, providing new insights for kidney development and function research.
Article
Computer Science, Information Systems
Yahya H. Yassin, Magnus Jahre, Per Gunnar Kjeldsberg, Snorre Aunet, Francky Catthoor
Summary: This paper presents a non-intrusive application controlled DVFS management implementation in the GEM5 simulator and a novel architecture independent energy model. By parametrizing and calibrating the energy model, and combining it with GEM5 output statistics, we successfully estimate the total energy consumption of the simulated system accurately.
JOURNAL OF SIGNAL PROCESSING SYSTEMS FOR SIGNAL IMAGE AND VIDEO TECHNOLOGY
(2021)
Article
Computer Science, Information Systems
Shikha Bhardwaj, Gitanjali Pandove, Pawan Kumar Dahiya
Summary: Research has shown that deep learning methods are more effective for image retrieval tasks compared to traditional machine learning algorithms. By comparing the performance of four different models on two benchmark datasets, the results show that the DBN-SBI model performs best in terms of precision, recall, and retrieval time.
JOURNAL OF INFORMATION SCIENCE AND ENGINEERING
(2021)
Article
Green & Sustainable Science & Technology
Hicham Karmouni, Mohamed Chouiekh, Saad Motahhir, Hassan Qjidaa, Mohamed Ouazzani Jamil, Mhamed Sayyouri
Summary: This paper proposes a fast and accurate algorithm for global maximum power point tracking in a photovoltaic system based on the sine-cosine algorithm. The algorithm demonstrates superior performance in partial shading conditions and high system efficiency, as validated and tested through simulations.
CLEANER ENGINEERING AND TECHNOLOGY
(2022)
Proceedings Paper
Computer Science, Information Systems
Fuga Nakamura, Ryosuke Harakawa, Masahiro Iwahashi
Summary: Product quantization (PQ) is a popular technique for fast image retrieval, but existing methods focus only on quantization errors. This paper proposes a novel PQ method that reduces the entropy of labels to improve retrieval performance, enabling fast and accurate retrieval when queries are given.
2021 ASIA-PACIFIC SIGNAL AND INFORMATION PROCESSING ASSOCIATION ANNUAL SUMMIT AND CONFERENCE (APSIPA ASC)
(2021)
Proceedings Paper
Engineering, Biomedical
Yu-Hsiang Lin, An-Cin Li, Ying-Ju Chen, Yuan Luo
Summary: A new method is proposed to improve the imaging speed and quality of differential phase contrast microscopy by modulating illumination pattern and pupil engineering, achieving accurate reconstruction of phase information under an isotropic phase transfer function and solving some imaging issues.
BIOMEDICAL IMAGING AND SENSING CONFERENCE 2021
(2021)
Article
Nanoscience & Nanotechnology
Lucien E. Weiss, Yael Shalev Ezra, Sarah Goldberg, Boris Ferdman, Omer Adir, Avi Schroeder, Onit Alalouf, Yoav Shechtman
NATURE NANOTECHNOLOGY
(2020)
Article
Computer Science, Artificial Intelligence
Elias Nehme, Boris Ferdman, Lucien E. Weiss, Tal Naor, Daniel Freedman, Tomer Michaeli, Yoav Shechtman
Summary: Fast acquisition of depth information is crucial for accurate 3D tracking, and wavefront coding can improve the precision of 3D tracking. Multi-channel wavefront coding has better performance in low-light applications and outperforms traditional single-channel designs.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2021)
Article
Chemistry, Physical
Alon Saguy, Tim N. Baldering, Lucien E. Weiss, Elias Nehme, Christos Karathanasis, Marina S. Dietz, Mike Heilemann, Yoav Shechtman
Summary: QAFKA is a fully automated workflow for quantitative analysis of SMLM data, extracting fluorophore blinking events and reporting the ratios of cluster types within the population. The algorithm accurately reports quantitative information on the monomer/dimer equilibrium of membrane receptors in single immobilized cells.
JOURNAL OF PHYSICAL CHEMISTRY B
(2021)
Article
Chemistry, Multidisciplinary
Nadav Opatovski, Yael Shalev Ezra, Lucien E. Weiss, Boris Ferdman, Reut Orange-Kedem, Yoav Shechtman
Summary: The study introduces a novel method for efficient multicolor 3D particle tracking over a large field of view, achieved through multiplexed point-spread-function engineering. The technique has been successfully demonstrated for tracking five types of emitters in vitro and colocalization of DNA loci in live yeast cells.
Article
Multidisciplinary Sciences
Reut Orange-Kedem, Elias Nehme, Lucien E. Weiss, Boris Ferdman, Onit Alalouf, Nadav Opatovski, Yoav Shechtman
Summary: Diffractive optical elements are widely used for reshaping wavefronts efficiently, and by using liquid immersion, authors enable microscale 3D-printed optics to behave like nanoscopic structures while maintaining high performance.
NATURE COMMUNICATIONS
(2021)
Article
Genetics & Heredity
Daniel Allen, Lucien E. Weiss, Alon Saguy, Michael Rosenberg, Ortal Iancu, Omri Matalon, Ciaran Lee, Katia Beider, Arnon Nagler, Yoav Shechtman, Ayal Hendel
Summary: CRISPR-Cas technology has brought revolutionary changes to gene editing, but concerns about off-target interactions and genotoxicity remain. The current practice of optimizing genome-editing parameters in preclinical studies is expensive and time-consuming. However, this study introduces a flow-based imaging method coupled with deep learning for image analysis to characterize DNA damage. The findings show that guide RNAs with higher genome-editing activity induce greater DNA damage response, even differentiating single on-target guide RNAs with varying levels of off-target interactions. This simplifies the process of evaluating and screening genome-editing parameters, enabling safer and more effective gene therapy applications.
Article
Multidisciplinary Sciences
Omer Adir, Mia R. Albalak, Ravit Abel, Lucien E. Weiss, Gal Chen, Amit Gruber, Oskar Staufer, Yaniv Kurman, Ido Kaminer, Jeny Shklover, Janna Shainsky-Roitman, Ilia Platzman, Lior Gepstein, Yoav Shechtman, Benjamin A. Horwitz, Avi Schroeder
Summary: In this study, the authors designed bioluminescent signaling mechanisms for intracellular and intercellular synthetic-to-natural cell communication, providing a new tool for controlling engineered processes inside tissues.
NATURE COMMUNICATIONS
(2022)
Article
Multidisciplinary Sciences
Tal Naor, Yevgeni Nogin, Elias Nehme, Boris Ferdman, Lucien E. Weiss, Onit Alalouf, Yoav Shechtman
Summary: The study of cell cycle progression and regulation is important for our understanding of fundamental biophysics, aging, and disease mechanisms. The researchers used high spatiotemporal resolution microscopy to investigate chromatin dynamics in mouse embryonic fibroblast cells and found varying levels of constraint on chromatin during different stages of the cell cycle.
Article
Cell Biology
Egor Sedov, Elle Koren, Sucheta Chopra, Roi Ankawa, Yahav Yosefzon, Marianna Yusupova, Lucien E. Weiss, Adnan Mahlys, Arad Soifer, Alona Feldman, Chen Luxenburg, Yoav Shechtman, Yaron Fuchs
Summary: In this study, we discovered that the GPI-anchored protein THY1 inhibits the activity of YAP in the skin through multiple molecular mechanisms. Loss of THY1 leads to increased adhesion, resulting in the dissociation of adherens junction complex and the release and translocation of YAP. Increased YAP-dependent proliferation in Thy1(-/-) mice enhances wound repair and hair follicle regeneration.
NATURE CELL BIOLOGY
(2022)
Article
Biochemical Research Methods
Alon Saguy, Onit Alalouf, Nadav Opatovski, Soohyen Jang, Mike Heilemann, Yoav Shechtman
Summary: Single-molecule localization microscopy (SMLM) has significantly improved spatial resolution in biological imaging, but has limitations in observing dynamics at high temporal resolution. In this study, we introduce DBlink, a deep-learning-based method that reconstructs super spatiotemporal resolution videos from SMLM data. DBlink utilizes a convolutional neural network combined with a bidirectional long short-term memory network to capture long-term dependencies between different frames. Experimental results demonstrate the effectiveness of DBlink in various scenarios, including simulated structures and live-cell dynamic SMLM. This advancement in super-resolution imaging of dynamic processes in living cells is crucial.
Proceedings Paper
Engineering, Electrical & Electronic
Reut Orange-Kedem, Elias Nehme, Lucien E. Weiss, Boris Ferdman, Onit Alalouf, Nadav Opatovski, Yoav Shechtman
2021 CONFERENCE ON LASERS AND ELECTRO-OPTICS EUROPE & EUROPEAN QUANTUM ELECTRONICS CONFERENCE (CLEO/EUROPE-EQEC)
(2021)
Article
Multidisciplinary Sciences
Racheli Gordon-Soffer, Lucien E. Weiss, Ran Eshel, Boris Ferdman, Elias Nehme, Moran Bercovici, Yoav Shechtman