A5101700312- COVID-19 diagnosis using AI
- AI in cancer detection
- Radiomics and Machine Learning in Medical Imaging
- Advancements in Semiconductor Devices and Circuit Design
- Integrated Circuits and Semiconductor Failure Analysis
- Adversarial Robustness in Machine Learning
- Semiconductor materials and devices
- Anomaly Detection Techniques and Applications
- Postharvest Quality and Shelf Life Management
- Advanced Malware Detection Techniques
- Essential Oils and Antimicrobial Activity
- Advanced Neural Network Applications
- Human Pose and Action Recognition
- Forensic Fingerprint Detection Methods
- Nanocomposite Films for Food Packaging
- Video Surveillance and Tracking Methods
- Business Process Modeling and Analysis
- Advanced Vision and Imaging
- Radiation Effects in Electronics
- High-Velocity Impact and Material Behavior
- Multi-Agent Systems and Negotiation
- Cell Image Analysis Techniques
- Service-Oriented Architecture and Web Services
- Advanced Image Processing Techniques
- Advanced Optical Sensing Technologies
Zhejiang Gongshang University
2016-2023
Nanjing Forestry University
2022
Georgia Institute of Technology
2010
Fast Gradient Sign Method is a well-known method for adversarial sample attack. New samples could be generated by adding small perturbation to input images, however the usually need users select themselves. This paper focuses on FGSM attack in face recognition scenario and empirically evaluate multiple factors terms of performance. The results demonstrate sensitive many factors, such as size perturbation, number iterations, granularity perturbation.
In this paper, we address the problem of breast caner detection from multi-view mammograms. We present a novel cross-view attention module (CvAM) which implicitly learns to focus on cancer-related local abnormal regions and highlighting salient features by exploring information among four views screening mammography exam, e.g. asymmetries between left right breasts lesion correspondence two same breast. More specifically, proposed CvAM calculates spatial maps based view different enhance...
The total ionizing dose irradiation effects are investigated in Si vertical diffused MOSFETs (VDMOSs) with different gate dielectrics including single SiO 2 layer and double 3 N 4 /SiO layer. Radiation-induced holes trapping is greater for than Dielectric oxidation temperature dependent TID also studied. Holes induced negative threshold voltage shift smaller at lower temperature. Gate bias during leads to <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"...
Research on single-person pose estimation based deep neural networks has recently witnessed progress in both accuracy and execution efficiency. However, multiperson is still a challenging topic, partially because the object regions are selected greedily from proposals via class-agnostic nonmaximum suppression (NMS), misalignment redundant detection yields inaccurate human poses. Therefore, we consider how to obtain optimal input under conditions which intermediate label information not...
The task of multimodal cancer detection is to determine the locations and categories lesions by using different imaging techniques, which one key research methods for diagnosis. Recently, deep learning-based object has made significant developments due its strength in semantic feature extraction nonlinear function fitting. However, remains challenging morphological differences lesions, interpatient variability, difficulty annotation, artifacts. In this survey, we mainly investigate over 150...
Regrown source/drain technology is highly appreciated in III-V MOSFET, to reduce the thermal budget induced by otherwise activation annealing required ion implantation technology, as well resistances doping epitaxial materials. While regrown facing problems such high parasitic capacitances and electric field at gate edge toward drain side, leading leakage current. To relieve current problem improve frequency performances, low-k Si3N4 nitride spacer was introduced InGaAs MOSFET. Experimental...
Microtubule (MT) dynamics quantification includes modeling of elongation, rapid shortening, and pauses. It indicates the effect cancer treatment drug paclitaxel because causes MTs to bundle, which will in turn inhibit successful mitosis cancerous cells. Thus, automatic MT analysis has been researched intensely it allows for faster evaluation potential treatments better understanding effects on a cell. However, most current literatures still use manual initialization. In this work, we propose...