A5100448048- Low-power high-performance VLSI design
- VLSI and FPGA Design Techniques
- Anomaly Detection Techniques and Applications
- Handwritten Text Recognition Techniques
- Parallel Computing and Optimization Techniques
- VLSI and Analog Circuit Testing
- Advanced Neural Network Applications
- Muscle activation and electromyography studies
- Web Data Mining and Analysis
- Embedded Systems Design Techniques
- Advanced Control Systems Design
- Extremum Seeking Control Systems
- Stroke Rehabilitation and Recovery
- Currency Recognition and Detection
- Adaptive Control of Nonlinear Systems
- Advanced Control Systems Optimization
- Rocket and propulsion systems research
- Prosthetics and Rehabilitation Robotics
- Topic Modeling
- Network Security and Intrusion Detection
- Advanced Image and Video Retrieval Techniques
- Advanced Computational Techniques and Applications
- Vehicle License Plate Recognition
- Sustainable Supply Chain Management
- Industrial Vision Systems and Defect Detection
Ritsumeikan University
2021-2025
Chengdu Normal University
2020-2024
Lanzhou Jiaotong University
2023-2024
Beckman Research Institute
2024
City of Hope
2024
Xi’an University of Posts and Telecommunications
2024
Ningxia Medical University
2023
Beijing Health Vocational College
2023
South China University of Technology
2023
Chinese University of Hong Kong
2021-2023
Currently, with the rapid development of deep learning, neural networks (DNNs) have been widely applied in various computer vision tasks. However, pursuit performance, advanced DNN models become more complex, which has led to a large memory footprint and high computation demands. As result, are difficult apply real time. To address these issues, model compression focus research. Furthermore, techniques play an important role deploying on edge devices. This study analyzed methods assist...
Abstract Upper gastrointestinal (GI) hemorrhage is a common clinical emergency worldwide. Endoscopic hemostasis the current first line treatment. Nevertheless, for patients with severe active bleeding and challenging anatomy, endoscopic management can be still difficult usually requires higher level of surgical expertise. A simple effective method therefore in acute demand. Herein hemostatic hybrid hydrogel comprised hyaluronic acid polyethylene glycol stabilized by thiourea–catechol...
Androgen deprivation therapy (ADT) targeting androgen/androgen receptor (AR)- signaling pathways is the main for advanced prostate cancer (PCa). However, ADT eventually fails in most patients who consequently develop castration-resistant (CRPC). While more potent AR antagonists and blockers androgen synthesis were developed to improve clinical outcomes, they also show induce diverse CRPC phenotypes. Specifically, AR- neuroendocrine-null PCa, DNPC, occurs abiraterone enzalutamide-treated...
The traditional purpose of physical synthesis is to perform timing closure , i.e., create a placed design that meets its specifications while also satisfying electrical, routability, and signal integrity constraints. In modern flows, tools hardly ever achieve this goal in their first iteration. team must iterate by studying the output run, then potentially massage input, e.g., changing floorplan, assertions, pin locations, logic structures, etc., order hopefully better solution for next...
Protecting the skin from UV light irradiation in wet and underwater environments is challenging due to weak adhesion of existing sunscreen materials but highly desired. Herein we report a polyethyleneimine/thioctic acid/titanium dioxide (PEI/TA/TiO2) coacervate-derived hydrogel with robust, asymmetric, reversible bioadhesion effective UV-light-shielding ability. The PEI/TA/TiO2 complex coacervate can be easily obtained by mixing PEI solution TA/TiO2 powder. fluid deposited on spread into...
To solve the labor shortage, robots have dramatically changed world by combining powerful deep learning (DL) technology. Certainly, DL technology has become key point of widespread robot application. Efficient models depend on high-quality datasets and optimized architectures. However, some open contain anomalous data that degrade model performance. Moreover, complex structures high computational costs limit adoption models. This study proposes a dataset purification-based lightweight...
As very large-scale integration technology shrinks to fewer tracks per standard cell, e.g., from 10 7.5-track libraries (and lesser for 7 nm), there has been a rapid increase in the usage of multiple-row cells like two- and three-row flip-flops, buffers, etc., design closure. Additionally, multibit flip-flops or flop trays save power creates large that further complicate critical tasks, such as placement. Detailed placement happens be key optimization transform, which is repeatedly invoked...
Recently, handwritten Chinese character recognition has become an important research field in computer vision. With the development of deep learning, convolutional neural networks (CNNs) have demonstrated excellent performance However, CNNs are typically designed manually, which requires extensive experience and may lead to redundant computations. To solve these problems, this study, particle swarm optimization approach is incorporated into design a CNN for recognition, reducing computations...
The Transformer shows good prospects in computer vision. However, the Swin model has disadvantage of a large number parameters and high computational effort. To effectively solve these problems model, simplified (S-Swin Transformer) was proposed this article for handwritten Chinese character recognition. simplifies initial four hierarchical stages into three stages. In addition, new increases size window attention; patches is larger; perceptual field increased. As network deepens, becomes...
Delay faults are an increasingly important test challenge. Modeling bridge as delay helps tests to detect more faults. Traditional fault models incomplete because these only model the logic or not efficient use in for large circuits. In this article, we propose a physically realistic yet economical resistive well An accurate simple calculation method is proposed. We also enumerate all possible behaviors and present relationship between input patterns output behaviors, which useful ATPG. Our...
This paper presents a reversible data hiding scheme. The proposed scheme is based on the difference histogram shifting to spare space for hiding. Nine basic scan paths are defined, and this means all-directional adjacent pixel differences can be obtained. Due fact that grayscale values of pixels close each other, contains large number points with equal values. Hence, more embedded into cover image than previous works shifting. Furthermore, multi-layer embedding used increase capacity. In...
The researches on saliency detection for common images have been thoroughly explored in recent years, but little attention has paid to hyperspectral images. However, salient region can contribute finding significant targets In this paper, spectral is introduced the domain detecting images, which are computed by expanding Itti's visual model. Firstly, spatial and features extracted from image. conspicuity feature of first three principal components, orientations, angle visible band opponent...
Research shows that it is practical for the normal human movement mechanism to assist patients with stroke in robot-assisted gait rehabilitation. In passive training, effect of rehabilitation training can be improved by imitating walking. To make lower limb exoskeleton robot (LLER) move like a human, tracking control scheme based on data proposed this paper. The real obtained from healthy subjects using three-dimensional motion capture platform (3DMCP). Furthermore, characteristics are...
Real-world datasets usually are class-imbalanced and corrupted by label noise. To solve the joint issue of long-tailed distribution noise, most previous works aim to design a noise detector distinguish noisy clean samples. Despite their effectiveness, they may be limited in handling effectively unified way. In this work, we develop novel pseudo labeling method using class prototypes from perspective matching, which can solved with optimal transport (OT). By setting manually-specific...
Abstract With the wide application of deep learning in anomaly detection (AD), industrial vision AD has achieved remarkable success. However, current usually focuses on localization and rarely investigates classification. Furthermore, classification is currently requested for quality management reason analysis. Therefore, it essential to classify anomalies while improving accuracy AD. This paper designs a novel multi-classification (MCAD) framework achieve high-accuracy with an function. In...
We propose a new yield estimation algorithm which estimates the acceptability region as union of spherical cones. The works by dividing input parameter space into approximately equi-probable cones, efficiently estimating refined weight contributions for each cone, then combining results to get total yield. is broadly similar worst-case-distances method, but more generally applicable cases with -for example- multiple failure regions. quite accurate, and offers several orders (>100x) magnitude...