Article
Biochemical Research Methods
Xiaowen Feng, Haoyu Cheng, Daniel Portik, Heng Li
Summary: hifiasm-meta is a software tool designed for assembling metagenomes using high-accuracy long-read data, which can reconstruct bacterial genomes in microbial communities more accurately.
Article
Biochemical Research Methods
Victor Rodriguez-lopez, Wilson Geisler, Carlos Dorronsoro
Summary: Tunable lenses enable the measurement of visual sensitivity to rapid changes in optical power, surpassing mechanical limits. Using this system, the spatiotemporal defocus sensitivity function (STDSF) and the limits of human defocus perception were measured for the first time.
BIOMEDICAL OPTICS EXPRESS
(2023)
Article
Optics
Gang Wen, Simin Li, Linbo Wang, Xiaohu Chen, Zhenglong Sun, Yong Liang, Xin Jin, Yifan Xing, Yaming Jiu, Yuguo Tang, Hui Li
Summary: HiFi-SIM is a high-fidelity SIM reconstruction algorithm that can reduce artifacts and improve axial sectioning without loss of fine structures, making it a daily imaging tool with lower calibration and adjustment requirements.
LIGHT-SCIENCE & APPLICATIONS
(2021)
Article
Multidisciplinary Sciences
Wei Chen, Ryan G. Natan, Yuhan Yang, Shih-Wei Chou, Qinrong Zhang, Ehud Y. Isacoff, Na Ji
Summary: Studying neuronal activity at synapses requires high spatiotemporal resolution, which can be achieved through adaptive optics to correct sample-induced aberrations at depth for high spatial resolution. The combination of Bessel focus with two-photon fluorescence microscopy allows for fast volumetric imaging at subcellular lateral resolution. The developed AO method corrects distorted wavefront of Bessel focus at the objective focal plane, demonstrating significant improvements in sensitivity and resolution of structural and functional measurements of synapses in vivo.
NATURE COMMUNICATIONS
(2021)
Article
Computer Science, Information Systems
Zongyi Xu, Wei Chang, Yindi Zhu, Le Dong, Huiyu Zhou, Qianni Zhang
Summary: Two approaches are proposed to estimate high-fidelity human body models, one based on 3D scanner point clouds and the other based on 2D images from smartphones. Extensive experiments demonstrate that both approaches can robustly build believable and animatable human body models.
IEEE TRANSACTIONS ON MULTIMEDIA
(2021)
Article
Computer Science, Artificial Intelligence
Shu Chen, Yaxin Xu, Zhengdong Pu, Jianquan Ouyang, Beiji Zou
Summary: SkeletonPose is a method that improves the accuracy of 3D human pose estimation by using human skeleton constraints. By combining data-driven and calculation methods, the proposed approach regresses the z-coordinate of the root joint using deep convolutional networks and calculates the 3D human pose based on the skeleton length invariance constraint. This method reduces pose estimation errors by considering the skeleton length prior. Evaluation results show that SkeletonPose achieves better performance compared to other state-of-the-art pose estimation approaches.
KNOWLEDGE-BASED SYSTEMS
(2022)
Article
Mechanics
Sudeepta Mondal, Soumalya Sarkar
Summary: Data-driven prediction of spatiotemporal fields in fluid flow problems has become increasingly important. However, the accuracy of prevalent approaches is often affected by the scarcity of data, especially when generating high-fidelity data is expensive. This article proposes a novel multi-fidelity spatiotemporal modeling approach to reduce the overhead of high-fidelity simulations and improve the accuracy of predictions.
Article
Engineering, Marine
Yichen Huang, Jing Li, Bin Xie, Zhiliang Lin, Alistair G. L. Borthwick
Summary: A novel high-fidelity numerical model based on the Navier-Stokes equations is proposed to simulate solitary wave propagation, accurately predicting the characteristics of solitary waves and reducing undesirable numerical effects. The model's performance is validated through comparisons with analytical solutions, alternative numerical results, and experimental data.
Article
Computer Science, Software Engineering
Fukang Liu, Mingwen Shao, Fan Wang, Lixu Zhang
Summary: Generative Adversarial Network (GAN) inversion is used to invert an input image into the latent space of a pre-trained GAN, allowing for downstream visual tasks. To address the limitations of deep neural networks in learning high-frequency image distributions, we propose a high-frequency information generation guidance branch for more accurate image reconstruction. By utilizing the frequency domain difference between the initial reconstructed image and the original image, we guide the generator to reconstruct the high-frequency details of challenging images. Additionally, we introduce focal frequency loss to expedite the learning process of previously unlearned high-frequency information. Extensive experiments on benchmark datasets demonstrate the superiority of our method and the faster convergence of the network.
COMPUTERS & GRAPHICS-UK
(2023)
Article
Engineering, Marine
Jihao Fan, Wenyang Duan, Limin Huang, Lu Zhang, Ke Yang
Summary: Research on the flow characteristics of incompressible fluids in ship and ocean engineering is important. Direct numerical simulation method is accurate but computationally expensive, while the Reynold-averaged Navier-Stokes method is widely used due to its higher computational efficiency but lower accuracy. Deep learning has provided a new approach for studying fluid flow characteristics, but current models have limitations. To address these issues, a high-fidelity flow field reconstruction model, named HFRIF, is proposed.
Article
Biochemistry & Molecular Biology
David Fawkner-Corbett, Agne Antanaviciute, Kaushal Parikh, Marta Jagielowicz, Ana Sousa Geros, Tarun Gupta, Neil Ashley, Doran Khamis, Darren Fowler, Edward Morrissey, Chris Cunningham, Paul R. Johnson, Hashem Koohy, Alison Simmons
Summary: This study uses single-cell RNA sequencing and spatial transcriptomics to investigate human intestinal development, identifying various cell states, developmental programs, and differentiation hierarchies. It provides insights into the formation of different cell types in the gut and offers valuable resources for further research in this field.
Article
Optics
Ning Xu, Guoxuan Liu, Qiaofeng Tan
Summary: Far-field super-resolution microscopy has been made possible with the use of a 2D multilevel diffractive optical element (DOE) and a modified deconvolution algorithm, enabling high-resolution and high-fidelity imaging at the nanoscale.
LASER & PHOTONICS REVIEWS
(2022)
Article
Biology
Abdelmoneim Eshra, Hartmut Schmidt, Jens Eilers, Stefan Hallermann
Summary: The basal free Ca2+ concentration critically controls action potential-evoked release, indicating a high-affinity Ca2+ sensor for vesicle priming. There is a surprisingly shallow and non-saturating relationship between release rate and intracellular Ca2+ concentration up to 50 μM. The rate of vesicle replenishment during sustained elevated intracellular Ca2+ concentration exhibited little Ca2+-dependence.
Article
Biotechnology & Applied Microbiology
Sabrina Hepner, Konstantin Kuleshov, Ave Tooming-Kunderud, Nikolas Alig, Alexander Gofton, Sherwood Casjens, Robert E. Rollins, Alexandra Dangel, Evangelos Mourkas, Samuel K. Sheppard, Andreas Wieser, Johannes Huebner, Andreas Sing, Volker Fingerle, Gabriele Margos
Summary: This study developed a workflow pipeline using high-fidelity PacBio sequencing to obtain complete and high quality Borrelia genome assemblies. The results demonstrated that the pipeline successfully resolved both chromosomal and plasmid sequences.
Article
Multidisciplinary Sciences
Majid Bizhani, Omid Haeri Ardakani, Edward Little
Summary: Imaging methods have broad applications in geosciences, and using convolutional neural networks can rapidly process digital rock images, improving image quality.
SCIENTIFIC REPORTS
(2022)