13197 research outputs found
Sort by
HMA-Depth: A New Monocular Depth Estimation Model Using Hierarchical Multi-Scale Attention
Get PDFMonocular depth estimation is an essential technique for tasks like 3D reconstruction. Although many works have emerged in recent years, they can be improved by better utilizing the multi-scale information of the input images, which is proved to be one of the keys in generating high-quality depth estimations. In this paper, we propose a new monocular depth estimation method named HMA-Depth, in which we follow the encoder-decoder scheme and combine several techniques such as skip connections and the atrous spatial pyramid pooling. To obtain more precise local information from the image while keeping a good understanding of the global context, a hierarchical multi-scale attention module is adopted and its outputs are combined to generate the final output that is with both good details and good overall accuracy. Experimental results on two commonly-used datasets prove that HMA-Depth can outperform the existing approaches. Code is available 1 1 https://github.com/saranew/HMADepth.conference pape
NAIST English-to-Japanese Simultaneous Translation System for IWSLT 2021 Simultaneous Text-to-text Task
No full textThis paper describes NAIST’s system for the English-to-Japanese Simultaneous Text-to-text Translation Task in IWSLT 2021 Evaluation Campaign. Our primary submission is based on wait-k neural machine translation with sequence-level knowledge distillation to encourage literal translation.conference pape
Leveraging IoT and Weather Conditions to Estimate the Riders Waiting for the Bus Transit on Campus
Get PDFThe communication technology revolution in this era has increased the use of smartphones in the world of transportation. In this paper, we propose to leverage IoT device data, capturing passengers' smartphones' Wi-Fi data in conjunction with weather conditions to predict the expected number of passengers waiting at a bus stop at a specific time using deep learning models. Our study collected data from the transit bus system at James Madison University (JMU) in Virginia, USA. This paper studies the correlation between the number of passengers waiting at bus stops and weather conditions. Empirically, an experiment with several bus stops in JMU, was utilized to confirm a high precision level. We compared our Deep Neural Network (DNN) model against two baseline models: Linear Regression (LR) and a Wide Neural Network (WNN). The gap between the baseline models and DNN was 35% and 14% better Mean Squared Error (MSE) scores for predictions in favor of the DNN compared to LR and WNN, respectively.conference pape
Studies on roles for nonsense-mediated mRNA decay in auxin response during seedling development and de novo shoot organogenesis
Get PDF奈良先端科学技術大学院大学博士(バイオサイエンス)doctoral thesi
Ultrasmall compact CMOS imaging system for bioluminescence reporter-based live gene expression analysis
No full textSignificance: Gene expression analysis is an important fundamental area of biomedical research. However, live gene expression imaging has proven challenging due to constraints in conventional optical devices and fluorescent reporters.Aim: Our aim is to develop smaller, more cost-effective, and versatile imaging capabilities compared with conventional devices. Bioluminescence reporter-based gene expression analysis was targeted due to its advantages over fluorescence-based imaging.Approach: We created a small compact imaging system using micro-CMOS image sensors (μCIS). The μCIS model had an improved pixel design and a patterned absorption filter array to detect the low light intensity of bioluminescence.Results: The device demonstrated lower dark current, lower temporal noise, and higher sensitivity compared with previous designs. The filter array enabled us to subtract dark current drift and attain a clearer light signal. These improvements allowed us to measure bioluminescence reporter-based gene expression in living mammalian cells.Conclusion: Using our μCIS system for bioluminescence imaging in the future, the device can be implanted in vivo for simultaneous gene expression imaging, behavioral analysis, and optogenetic modulation.journal articl
Meta-analysis of single-cell RNA-seq data reveals phenotypic switching of immune cells in severe COVID-19 patients
No full textSevere acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection has resulted in the global coronavirus disease 2019 (COVID-19) pandemic. Despite several single-cell RNA sequencing (RNA-seq) studies, conclusions cannot be reached owing to the small number of available samples and the differences in technology and tissue types used in the studies. To better understand the cellular landscape and disease severity in COVID-19, we performed a meta-analysis of publicly available single-cell RNA-seq data from peripheral blood and lung samples of COVID-19 patients with varying degrees of severity. Patients with severe disease showed increased numbers of M1 macrophages in lung tissue, while the number of M2 macrophages was depleted. Cellular profiling of the peripheral blood showed a marked increase of CD14+, CD16+ monocytes and a concomitant depletion of overall B cells and CD4+, CD8+ T cells in severe patients when compared with moderate patients. Our analysis indicates the presence of faulty innate-to-adaptive switching, marked by a prolonged innate immune response and a dysregulated adaptive immune response in severe COVID-19 patients. Furthermore, we identified cell types with a transcriptome signature that can be used as a prognostic biomarker for disease state prediction and the effective therapeutic management of COVID-19 patients.journal articl
Filopodium-derived vesicles produced by MIM enhance the migration of recipient cells
Get PDFExtracellular vesicles (EVs) are classified as large EVs (l-EVs, or microvesicles) and small EVs (s-EVs, or exosomes). S-EVs are thought to be generated from endosomes through a process that mainly depends on the ESCRT protein complex, including ALG-2 interacting protein X (ALIX). However, the mechanisms of l-EV generation from the plasma membrane have not been identified. Membrane curvatures are generated by the bin-amphiphysin-rvs (BAR) family proteins, among which the inverse BAR (I-BAR) proteins are involved in filopodial protrusions. Here, we show that the I-BAR proteins, including missing in metastasis (MIM), generate l-EVs by scission of filopodia. Interestingly, MIM-containing l-EV production was promoted by in vivo equivalent external forces and by the suppression of ALIX, suggesting an alternative mechanism of vesicle formation to s-EVs. The MIM-dependent l-EVs contained lysophospholipids and proteins, including IRS4 and Rac1, which stimulated the migration of recipient cells through lamellipodia formation. Thus, these filopodia-dependent l-EVs, which we named as filopodia-derived vesicles (FDVs), modify cellular behavior.journal articl
Promotion and Upregulation of a Plasma Membrane Proton-ATPase Strategy: Principles and Applications
No full textPlasma membrane proton-ATPase (PM H+-ATPase) is a primary H+ transporter that consumes ATP in vivo and is a limiting factor in the blue light-induced stomatal opening signaling pathway. It was recently reported that manipulation of PM H+-ATPase in stomatal guard cells and other tissues greatly improved leaf photosynthesis and plant growth. In this report, we review and discuss the function of PM H+-ATPase in the context of the promotion and upregulation H+-ATPase strategy, including associated principles pertaining to enhanced stomatal opening, environmental plasticity, and potential applications in crops and nanotechnology. We highlight the great potential of the promotion and upregulation H+-ATPase strategy, and explain why it may be applied in many crops in the future.journal articl
Software-Hardware Co-design and its Evaluation Framework for Decimal Multiplication
Get PDF奈良先端科学技術大学院大学博士(工学)doctoral thesi