Article
Biology
Zhichao Liu, Luhong Jin, Jincheng Chen, Qiuyu Fang, Sergey Ablameyko, Zhaozheng Yin, Yingke Xu
Summary: Advanced microscopy allows for acquisition of time-lapse images to visualize dynamic characteristics, but requires processing with complex algorithms. Deep learning technologies, particularly convolutional neural networks, have been increasingly applied in bioimage processing with inspiring outcomes. Challenges remain in training dataset acquisition and evaluation for the field.
COMPUTERS IN BIOLOGY AND MEDICINE
(2021)
Article
Computer Science, Artificial Intelligence
Sehyung Lee, Hideaki Kume, Hidetoshi Urakubo, Haruo Kasai, Shin Ishii
Summary: The paper proposes a novel approach for restoring the image quality of three-dimensional neural imaging using convolutional neural networks. By estimating and fusing the intersection regions of images captured from three orthogonal viewpoints, the method improves the image quality along the depth direction and enhances the identification of neural connectivity.
Review
Computer Science, Artificial Intelligence
Jinghua Zhang, Chen Li, Yimin Yin, Jiawei Zhang, Marcin Grzegorzek
Summary: Microorganisms are widely distributed in human daily living environment and play essential roles in environmental pollution control and health. Automatic microorganism image analysis faces challenges such as the need for robust algorithms and addressing image characteristic issues.
ARTIFICIAL INTELLIGENCE REVIEW
(2023)
Article
Computer Science, Artificial Intelligence
Gao Huang, Zhuang Liu, Geoff Pleiss, Laurens van der Maaten, Kilian Q. Weinberger
Summary: Recent work has shown that adding shorter connections in convolutional networks can make the network deeper, more accurate, and more efficient in training. This paper introduces Dense Convolutional Network (DenseNet), which connects each layer to every other layer in a feed-forward manner. DenseNets alleviate the vanishing-gradient problem, encourage feature reuse, and improve parameter efficiency, leading to significant improvements in object recognition tasks.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2022)
Article
Agriculture, Multidisciplinary
Liang Gong, Xiaofeng Du, Kai Zhu, Chenghui Lin, Ke Lin, Tao Wang, Qiaojun Lou, Zheng Yuan, Guoqiang Huang, Chengliang Liu
Summary: The study of plant growth state relies on root architecture parameters, with root segmentation being crucial to measuring these parameters. A new method based on a convolutional neural network was proposed for pixel-level segmentation of rice roots under strong noise, achieving an intersection over union (IoU) of 87.4%. This approach provides an automatic and fast pixel-level root segmentation method, essential for root morphology analysis.
COMPUTERS AND ELECTRONICS IN AGRICULTURE
(2021)
Article
Plant Sciences
Carola Figueroa-Flores, Pablo San-Martin
Summary: This study evaluates the performance of several Deep Learning models for classifying images of Chilean native flora and highlights their potential in accurately classifying these images. The results contribute to enhancing the understanding of Chilean plant species and fostering awareness among the general public.
FRONTIERS IN PLANT SCIENCE
(2023)
Article
Computer Science, Interdisciplinary Applications
Liang Han, Hang Su, Zhaozheng Yin
Summary: Phase contrast microscopy is a noninvasive imaging technique that can monitor the behavior of transparent cells without staining or altering them. However, the imaging images of phase contrast microscopy contain artifacts that hinder cell segmentation and detection. In this research, we accurately formulated the imaging model of phase contrast microscopy and proposed an image restoration procedure using a deep neural network, which enables high quality cell segmentation.
IEEE TRANSACTIONS ON MEDICAL IMAGING
(2023)
Article
Environmental Sciences
Jiangbo Xi, Okan K. Ersoy, Ming Cong, Chaoying Zhao, Wei Qu, Tianjun Wu
Summary: This paper proposes a wide and deep Fourier network for efficient feature learning in hyperspectral remote sensing image (HSI) classification. The method utilizes pruned features extracted in the frequency domain to extract hierarchical features layer-by-layer. Experimental results show that the proposed method achieves excellent performance in HSI classification, with the ability to be implemented in lightweight embedded computing platforms.
Article
Mechanics
Joao M. Machado, Joao Manuel R. S. M. Trvares, Pedro P. Camanho, Nuno Correia
Summary: This study proposes a machine-learning approach based on a convolutional neural network architecture to automatically parse the void content of optical microscopy images without parameter tuning. Experimental results show that this approach accurately parses void content from microscopy images, outperforming traditional thresholding algorithms.
COMPOSITE STRUCTURES
(2022)
Article
Ecology
Arata Yabuki, Hidetoshi Ikeno, Masako Dannoura
Summary: In this study, we attempted to automate the analysis of fine root images using convolutional neural network, and we successfully extracted fine roots using our software. This software enables the automatic processing of scanned images, accelerating the study of fine root dynamics.
METHODS IN ECOLOGY AND EVOLUTION
(2022)
Article
Biology
Zhaoshan Liu, Qiujie Lv, Ziduo Yang, Yifan Li, Chau Hung Lee, Lei Shen
Summary: This review summarizes the core component of the transformer, the attention mechanism, and its detailed structures, as well as the recent progress of the transformer in the field of medical image analysis. The experiments conducted in this review demonstrate that transformer-based methods outperform existing methods according to multiple evaluation metrics. Finally, the open challenges and future opportunities in this field are discussed.
COMPUTERS IN BIOLOGY AND MEDICINE
(2023)
Article
Computer Science, Artificial Intelligence
Shiye Lei, Fengxiang He, Yancheng Yuan, Dacheng Tao
Summary: This article finds that neural networks with less variability in decision boundaries have better generalizability. The experiments show significant negative correlations between decision boundary variability and generalizability. The article introduces the concepts of algorithm DB variability and (epsilon, eta)-data DB variability to measure variability in decision boundaries.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Ronghua Shang, Jiaming Wang, Licheng Jiao, Xiaohui Yang, Yangyang Li
Summary: This paper proposes a spatial feature-based convolutional neural network (SF-CNN) for solving PolSAR classification problems. The special structure of SF-CNN can expand the training set by combining different samples and enhance the network's ability to extract discriminative features in low-dimensional feature space. Experimental results show that SF-CNN outperforms standard CNN in PolSAR image classification tasks.
APPLIED SOFT COMPUTING
(2022)
Article
Environmental Sciences
Jiangbo Xi, Ming Cong, Okan K. Ersoy, Weibao Zou, Chaoying Zhao, Zhenhong Li, Junkai Gu, Tianjun Wu
Summary: The paper proposed a dynamic wide and deep neural network (DWDNN) for HSI classification, including multiple efficient wide sliding window and subsampling (EWSWS) networks that can grow dynamically with problem complexity. Compared to other deep learning methods, the proposed approach achieved the highest test accuracies.
Article
Chemistry, Multidisciplinary
Haoyang Ni, Zhenyao Wu, Xinyi Wu, Jacob G. Smith, Michael J. Zachman, Jian-min Zuo, Lili Ju, Guannan Zhang, Miaofang Chi
Summary: The atomic configurations of atomically dispersed catalysts (ADCs), such as atom-atom distances and clustering, greatly affect their catalytic performance. This study presents a CNN-based algorithm that can quantify the spatial arrangement of different adatom configurations. The algorithm was proven effective in accurately identifying atom positions and analyzing large data sets of ADCs. It offers a robust method to overcome the bottleneck in STEM analysis for ADC catalyst research and has the potential to be used as an on-the-fly analysis tool for catalysts in future in situ microscopy experiments.
Article
Multidisciplinary Sciences
Bianca Dumitrascu, Soledad Villar, Dustin G. Mixon, Barbara E. Engelhardt
Summary: Single-cell technologies allow for characterization of complex cell populations at unprecedented scale and resolution. The method proposed in this study uses linear programming for supervised genetic marker selection and provides a Python package scGeneFit for implementation.
NATURE COMMUNICATIONS
(2021)
Article
Biochemical Research Methods
Jonathan Lu, Bianca Dumitrascu, Ian C. McDowell, Brian Jo, Alejandro Barrera, Linda K. Hong, Sarah M. Leichter, Timothy E. Reddy, Barbara E. Engelhardt
Summary: BETS is a method for inferring causal gene networks from gene expression time series, its efficiency and parallelization allow for quick analysis of large datasets and competitive performance in benchmark testing. Through external data validation, BETS can accurately infer activating or inhibitory causal effects.
PLOS COMPUTATIONAL BIOLOGY
(2021)
Article
Cell Biology
Jamie McGinn, Adrien Hallou, Seungmin Han, Kata Krizic, Svetlana Ulyanchenko, Ramiro Iglesias-Bartolome, Frances J. England, Christophe Verstreken, Kevin J. Chalut, Kim B. Jensen, Benjamin D. Simons, Maria P. Alcolea
Summary: Research shows that mechanical stretch in the developing oesophagus promotes the emergence of a specific basal cell population, facilitating the transition from physiological strain to adult homeostasis. During the transition period, mechanical strain accumulates at the organ level, guiding the development of oesophageal epithelial cells towards adult homeostasis.
NATURE CELL BIOLOGY
(2021)
Editorial Material
Biochemical Research Methods
Ugis Sarkans, Wah Chiu, Lucy Collinson, Michele C. Darrow, Jan Ellenberg, David Grunwald, Jean-Karim Heriche, Andrii Iudin, Gabriel G. Martins, Terry Meehan, Kedar Narayan, Ardan Patwardhan, Matthew Robert Geoffrey Russell, Helen R. Saibil, Caterina Strambio-De-Castillia, Jason R. Swedlow, Christian Tischer, Virginie Uhlmann, Paul Verkade, Mary Barlow, Omer Bayraktar, Ewan Birney, Cesare Catavitello, Christopher Cawthorne, Stephan Wagner-Conrad, Elizabeth Duke, Perrine Paul-Gilloteaux, Emmanuel Gustin, Maria Harkiolaki, Pasi Kankaanpaa, Thomas Lemberger, Jo McEntyre, Josh Moore, Andrew W. Nicholls, Shuichi Onami, Helen Parkinson, Maddy Parsons, Marina Romanchikova, Nicholas Sofroniew, Jim Swoger, Nadine Utz, Lenard M. Voortman, Frances Wong, Peijun Zhang, Gerard J. Kleywegt, Alvis Brazma
Summary: The study proposes draft metadata guidelines to address the needs of diverse communities within light and electron microscopy, with the hope of stimulating discussions about their implementation and future extension.
Article
Biochemical Research Methods
Virginie Uhlmann, Zsuzsanna Puspoki, Adrien Depeursinge, Michael Unser, Daniel Sage, Julien Fageot
Summary: This article introduces Steer'n'Detect, an ImageJ plugin that implements a recently published algorithm for detecting patterns of interest at any orientation from a single template in 2D images with high accuracy. Compared to traditional template matching methods, Steer'n'Detect performs better in terms of speed and robustness, while ensuring accurate results even in the presence of noise.
Article
Engineering, Electrical & Electronic
Virginie Uhlmann, Laurene Donati, Daniel Sage
Summary: This article discusses the challenges researchers face when using deep learning models in bioimaging applications and provides good practices to address these challenges. It aims to foster discussions around guidelines for the appropriate deployment of deep learning in the field of biology, with the goal of accelerating the adoption of novel deep learning technologies.
IEEE SIGNAL PROCESSING MAGAZINE
(2022)
Article
Multidisciplinary Sciences
Luca Rosalia, Adrien Hallou, Laurence Cochrane, Thierry Savin
Summary: The mechanical properties of soft biological tissues have important implications in physiology and disease, influencing cell behavior and contributing to tissue development, maintenance, and repair. However, existing tools have limitations that hinder a comprehensive understanding of soft tissue biomechanics. In this study, an instrument based on closed-loop interaction between an electromagnetic actuator and an optical strain sensor is developed for high-fidelity uniaxial tensile testing of soft biological tissues in controlled environmental conditions. The instrument's enhanced reliability makes it an ideal platform for future studies on soft tissue mechanics.
Article
Biochemical Research Methods
Ethan A. G. Baker, Denis Schapiro, Bianca Dumitrascu, Sanja Vickovic, Aviv Regev
Summary: As the spatially resolved multiplex profiling of RNA and proteins becomes more prominent, understanding the statistical power for testing specific hypotheses in such experiments is crucial. This study introduces a method for generating tunable in silico tissues and constructs a computational framework for spatial power analysis using spatial profiling data sets. The framework can be applied to diverse spatial data modalities and tissues of interest, providing insights for spatial omics studies.
Article
Oncology
Elodie Grockowiak, Claudia Korn, Justyna Rak, Veronika Lysenko, Adrien Hallou, Francesca M. Panvini, Matthew Williams, Claire Fielding, Zijian Fang, Eman Khatib-Massalha, Andres Garcia-Garcia, Juan Li, Reema A. Khorshed, Sara Gonzalez-Anton, E. Joanna Baxter, Anjali Kusumbe, Bridget S. Wilkins, Anna Green, Benjamin D. Simons, Claire N. Harrison, Anthony R. Green, Cristina Lo Celso, Alexandre P. A. Theocharides, Simon Mendez-Ferrer
Summary: Aging promotes the growth of hematopoietic stem cells (HSCs) with somatic mutations related to myeloid malignancies, such as myeloproliferative neoplasms (MPNs). Different bone marrow niches were found for HSCs in different MPN subtypes. JAK-STAT signaling regulates HSC polarity, niche interaction, and mutant cell expansion. The interactions between HSCs and niches influence the expansion rate and therapy response of cells with the same clonal hematopoiesis oncogenic driver.
Article
Computer Science, Interdisciplinary Applications
James Klatzow, Giovanni Dalmasso, Neus Martinez-Abadias, James Sharpe, Virginie Uhlmann
Summary: Modern microscopy technologies allow for 3D imaging of biological objects, enabling quantitative assessment of morphology. The mu Match pipeline introduces a state-of-the-art shape correspondence algorithm for soft-tissue objects without the need for landmarks, establishing correspondence in a fully automated manner.
FRONTIERS IN COMPUTER SCIENCE
(2022)
Proceedings Paper
Engineering, Biomedical
Ethan Cohen, Virginie Uhlmann
Summary: AURA-net is a convolutional neural network designed for segmenting phase-contrast microscopy images, utilizing transfer learning and Attention mechanisms to enhance training efficiency, suitable for small datasets. Additionally, it employs a loss function inspired by active contours to further improve performance.
2021 IEEE 18TH INTERNATIONAL SYMPOSIUM ON BIOMEDICAL IMAGING (ISBI)
(2021)
Article
Computer Science, Artificial Intelligence
Julien Fageot, Virginie Uhlmann, Zsuzsanna Puespoeki, Benjamin Beck, Michael Unser, Adrien Depeursinge
Summary: A pipeline for pattern detection in images is presented, utilizing a continuous-domain additive image model and optimal filter computation method. The approach involves discretization on polar grids and improving detection performance by exploiting the power-spectrum decay of background statistics.
IEEE TRANSACTIONS ON IMAGE PROCESSING
(2021)