A5102016617- Nuclear reactor physics and engineering
- Semiconductor materials and devices
- Nuclear Materials and Properties
- Advancements in Semiconductor Devices and Circuit Design
- Generative Adversarial Networks and Image Synthesis
- Plasma and Flow Control in Aerodynamics
- Nuclear Physics and Applications
- Nuclear and radioactivity studies
- Advanced Vision and Imaging
- Integrated Circuits and Semiconductor Failure Analysis
- Human Pose and Action Recognition
- Solid State Laser Technologies
- Anomaly Detection Techniques and Applications
- Advanced Image Processing Techniques
- Advanced Neural Network Applications
- Synthetic Aperture Radar (SAR) Applications and Techniques
- Silicon and Solar Cell Technologies
- Aerosol Filtration and Electrostatic Precipitation
- Laser Design and Applications
- Lattice Boltzmann Simulation Studies
- Nuclear Engineering Thermal-Hydraulics
- Optical Systems and Laser Technology
- Image and Signal Denoising Methods
- Graphite, nuclear technology, radiation studies
- Laser and Thermal Forming Techniques
Ryukoku University
2024
Preferred Networks (Japan)
2017-2024
Mitchell Institute
2023
Texas A&M University
2023
Toho University
2023
National Police Academy
2022
Tokyo Institute of Technology
1998-2014
Tohoku University
2012
Kyushu University
2006-2010
Inpex (Japan)
2009
In this paper, we propose a generative model, Temporal Generative Adversarial Nets (TGAN), which can learn semantic representation of unlabeled videos, and is capable generating videos. Unlike existing (GAN)-based methods that generate videos with single generator consisting 3D deconvolutional layers, our model exploits two different types generators: temporal an image generator. The takes latent variable as input outputs set variables, each corresponds to frame in video. transforms such...
Despite significant progress of deep learning in the field computer vision, there has not been a software library that covers these methods unifying manner. We introduce ChainerCV, is intended to fill this gap. ChainerCV supports numerous neural network models as well components needed conduct research vision. These implementations emphasize simplicity, flexibility and good engineering practices. The designed perform on par with results reported published papers its tools can be used...
Despeckling is a crucial noise reduction task in improving the quality of synthetic aperture radar (SAR) images. Directly obtaining noise-free SAR images challenging that has hindered development accurate despeckling algorithms. The advent deep learning facilitated study denoising models learn from only noisy However, existing methods deal solely with single-polarization and cannot handle multipolarization captured by modern satellites. In this work, we present an extension model for...
Despite significant progress of deep learning in the field computer vision, there has not been a software library that covers these methods unifying manner. We introduce ChainerCV, is intended to fill this gap. ChainerCV supports numerous neural network models as well components needed conduct research vision. These implementations emphasize simplicity, flexibility and good engineering practices. The designed perform on par with results reported published papers its tools can be used...
In this paper, we propose a generative model, Temporal Generative Adversarial Nets (TGAN), which can learn semantic representation of unlabeled videos, and is capable generating videos. Unlike existing (GAN)-based methods that generate videos with single generator consisting 3D deconvolutional layers, our model exploits two different types generators: temporal an image generator. The takes latent variable as input outputs set variables, each corresponds to frame in video. transforms such...
Abstract On March 20, 2005, a large M JMA 7.0 earthquake occurred in the offshore area, west of Fukuoka prefecture, northern Kyushu, Japan. A series joint observations were carried out by teams from several universities Japan with aim investigating aftershock activity. Six online telemetered and 17 offline recording seismic stations installed on land around area immediately followed occurrence mainshock. Because aftershocks located mainly regions, we also 11 ocean bottom seismometers (OBSs)...
The mean field (MF) methods are an energy optimization method for Markov random fields (MRFs). These methods, which have their root in solid state physics, estimate the marginal density of each site MRF graph by iterative computation, similarly to loopy belief propagation (LBP). It appears that, being shadowed LBP, MF not been seriously considered computer vision community. This study investigates whether these useful practical problems, particularly MPM (Maximum Posterior Marginal)...
We propose a fluorine (F) treatment technique that is suitable for threshold voltage (Vth) modulation in p-channel metal–oxide–semiconductor field-effect transistors (PMOSFETs) with the atomic layer deposition (ALD) TiN/HfO2 gate structure. A work function near band edge achieved using F without degradation of hole mobility PMOSFETs. Vth shift value almost same regardless length. The barrier height attributable to corresponds closely shift. It found modulates effective function. No n-channel...
(2002). Experimental and Theoretical Study of the Residual Product Nuclide Yields in 100-2600 MeV Proton-Irradiated Thin Targets. Journal Nuclear Science Technology: Vol. 39, No. sup2, pp. 1175-1178.
For a coexisting and collaborative society that incorporates humans robots, the detection, tracking, recognition of human motion are indispensable techniques for robot to safely securely interact with humans. The present paper proposes tracking system using distributed network cameras placed in sizeable environment, such as street or town. Model-based is adopted this system, an asynchronous process invoked updating estimation each camera individually. A 2D distance map created by Fast...
We present resonator-enhanced frequency conversion of a master-oscillator fiber power-amplifier. 88%-efficient 2 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">nd</sup> -harmonic generation results in > 20 W at 532 nm. A low-loss Czochralski-grown BBO generates record-high, 12-W continuous-wave, 4 xmlns:xlink="http://www.w3.org/1999/xlink">th</sup> radiation 266
In this paper, we demonstrate a wet treatment for the HfSix/HfO2 gate stack of n-type metal oxide semiconductor field effect transistor (nMOSFET) fabricated by gate-last process in order to scale down electrical thickness at inversion state Tinv value and reduce leakage Jg. As result, succeeded scaling 1.41 nm without mobility or Jg degradation ozone–water-last treatment. We found that high-density interfacial layer (IFL) is formed owing treatment, Hf diffusion IFL suppressed, which was...
In recent years, research and hacking verification experiments have been conducted on cybersecurity countermeasures in vehicles. However, establishing digital forensics technology related to vehicle remains a problem for clarifying the facts, methods, effects when via an in-vehicle network actually occurs. Digital refers methods procedures handling information recorded devices or elsewhere without loss of evidentiary value, is regarded as important by various law enforcement agencies,...
Self-induced vibration is possibly caused by the semi-closed pipeline combining automatic pressure regulator with constant flow valve. Thus, a test was conducted on site to analyze generation of coupled oscillation between and direct-acting Further, investigation in order show function valve, obtain self-excited oscillation, study control measure. Consequently, when valve used near setting flow, working mechanism activates using pipe large can be induced. On other hand, smaller than that not...
We investigated the vertical stress impact on FinFETs using in-situ electrical measurements in a nano-indenter setup. found that of I <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">d</inf> for N and P-type increases longer gate lengths. According to mechanical simulations, if is applied sample surface, FinFET feels not only compressive but also non-negligible horizontal stress. Furthermore, we confirmed in-plane stresses have different values...
Sentiment Analysis is essential in the area of marketing or communication. In recent years, visual sentiment analysis that estimates how people feel when looking at an image has been widely studied along with progress machine learning. The main difficulty to bridge affective gap translating from features labels. SentiBank, a framework, solved this problem and achieved high accuracy emotion prediction by text-based approach through representation adjective noun pairs. However, there...
Despeckling is a crucial noise reduction task in improving the quality of synthetic aperture radar (SAR) images. Directly obtaining noise-free SAR images challenging that has hindered development accurate despeckling algorithms. The advent deep learning facilitated study denoising models learn from only noisy However, existing methods deal solely with single-polarization and cannot handle multi-polarization captured by modern satellites. In this work, we present an extension model for...