A5072851503- Advanced Malware Detection Techniques
- Adversarial Robustness in Machine Learning
- 3D Shape Modeling and Analysis
- Surface Treatment and Residual Stress
- Anomaly Detection Techniques and Applications
- Microstructure and Mechanical Properties of Steels
- CCD and CMOS Imaging Sensors
- Advanced Optical Imaging Technologies
- Advanced Optical Sensing Technologies
- Advanced Neural Network Applications
- AI in cancer detection
- Advanced Surface Polishing Techniques
- Medical Image Segmentation Techniques
- Retinal Imaging and Analysis
- Digital Imaging for Blood Diseases
Harbin University of Science and Technology
2022
Carnegie Mellon University
2019
Zhejiang University
2019
Deep learning (DL) models are inherently vulnerable to adversarial examples - maliciously crafted inputs trigger target DL misbehave which significantly hinders the application of in security-sensitive domains. Intensive research on has led an arms race between adversaries and defenders. Such plethora emerging attacks defenses raise many questions: Which more evasive, preprocessing-proof, or transferable? effective, utility-preserving, general? Are ensembles multiple robust than individuals?...
We propose a novel approach for 3D mesh reconstruction from multi-view images. Our method takes inspiration large models like LRM that use transformer-based triplane generator and Neural Radiance Field (NeRF) model trained on However, in our method, we introduce several important modifications allow us to significantly enhance quality. First of all, examine the original architecture find shortcomings. Subsequently, respective architecture, which lead improved image representation more...
Modern text-to-video synthesis models demonstrate coherent, photorealistic generation of complex videos from a text description. However, most existing lack fine-grained control over camera movement, which is critical for downstream applications related to content creation, visual effects, and 3D vision. Recently, new methods the ability generate with controllable poses these techniques leverage pre-trained U-Net-based diffusion that explicitly disentangle spatial temporal generation. Still,...
Background: The essence of the plastic deformation material to be processed during cutting process is movement dislocations. By increasing density dislocations, material’s flow stress inevitably increases by raising strain rate. Objectives: three regions are divided quantitatively as thermal activation stage (101-3×103), nonthermal (10-3-101), and dislocation damping (above 3×103). Analyzing chip metallographic diagram process, it believed that rate effect caused two reasons. Methods: areas...