A5027591323- Crystallization and Solubility Studies
- X-ray Diffraction in Crystallography
- Crystallography and molecular interactions
- Advanced Image and Video Retrieval Techniques
- Particle Detector Development and Performance
- Advanced Neural Network Applications
- Superconducting Materials and Applications
- Advanced Vision and Imaging
- Particle accelerators and beam dynamics
- Muon and positron interactions and applications
- Visual Attention and Saliency Detection
- Twentieth Century Scientific Developments
- Magnetic confinement fusion research
- Image Enhancement Techniques
- Advanced Electron Microscopy Techniques and Applications
- Photocathodes and Microchannel Plates
- Vacuum and Plasma Arcs
- Hemodynamic Monitoring and Therapy
- Auditing, Earnings Management, Governance
- Radiation Detection and Scintillator Technologies
- Graphite, nuclear technology, radiation studies
- Cardiac, Anesthesia and Surgical Outcomes
- COVID-19 diagnosis using AI
- Laser-Matter Interactions and Applications
- Ultrasound in Clinical Applications
Hôpital Sainte-Marguerite
2022
University of California, Davis
2019
University of California, San Diego
2014-2018
Universidad Católica Santo Domingo
2016
Osaka University
2016
Given the vast amounts of video available online, and recent breakthroughs in object detection with static images, offers a promising new frontier. However, motion blur compression artifacts cause substantial frame-level variability, even videos that appear smooth to eye. Additionally, datasets tend have sparsely annotated frames. We present framework for improving captures temporal context encourages consistency predictions. First, we train pseudo-labeler, is, domain-adapted convolutional...
Object proposals for detecting moving or static video objects need to address issues such as speed, memory complexity and temporal consistency. We propose an efficient Video Proposal (VOP) generation method show its efficacy in learning a better object detector A deep-learning based learned using the proposed VOP achieves state-of-the-art detection performance on Youtube-Objects dataset. further clustering of VOPs which can efficiently be used streaming fashion. As opposed applying per-frame...
A new experiment, called DeeMe, which is designed to search for μ−e conversions with a sensitivity of O(10−14), in preparation at the Japan Proton Accelerator Research Complex (J-PARC). It utilizes high-quality pulsed proton beam from Rapid Cycling Synchrotron J-PARC. The detector DeeMe must tolerate large pulses prompt charged particles whose instantaneous hit rate as 70 GHz mm−2 time width 200 ns, and detect single electron that arrives delayed timing on order microseconds. special wire...
After the introduction of FlowNet and large scale synthetic dataset Flying Chairs, we witnessed a rapid growth deep learning based optical flow estimation algorithms. However, most these algorithms rely on very network to learn both small motions, making them less efficient. They also process each frame individually for video like MPI Sintel without using temporally correlated information across frames. This paper presents pyramid structured that estimates from coarse fine. We use much...
DeeMe project is an experiment searching for muon to electron conversion (µe conversion) signal with single event sensitivity at level 10 -14 by using a pulsed proton beam from J-PARC Rapid Cycling Synchrotron (RCS).The µe process called charged-Lepton Flavor Violation.This essentially forbidden in the Standard Model of particle physics, therefore, this one clean processes study physics beyond Model.DeeMe proposed Materials and Life Science Experimental Facility (MLF) it will be conducted...
We analyze and evaluate different initialization methods for motion layers segmentation, a powerful mid-level vision tool. estimate 6-D affine models from the optical flow corresponding to over-segmentations. Over-segmentations can be either uniform rectangular blocks; or adaptive sized super-pixels; based on any other clustering methods. present performance analysis of segmentation algorithms video sequences discuss relative pros cons these in terms hardware implementation issues.
MuSIC is