A5014321050- Fatigue and fracture mechanics
- Structural Health Monitoring Techniques
- Structural Load-Bearing Analysis
- Advanced Vision and Imaging
- Advanced Neural Network Applications
- Remote-Sensing Image Classification
- Model Reduction and Neural Networks
- Image Enhancement Techniques
- Advanced Image Processing Techniques
- Advanced Numerical Analysis Techniques
- Remote Sensing and Land Use
- Infrared Target Detection Methodologies
- Advanced Image and Video Retrieval Techniques
- Radiomics and Machine Learning in Medical Imaging
- Soil Mechanics and Vehicle Dynamics
- Multimodal Machine Learning Applications
- Non-Destructive Testing Techniques
- Neural Networks and Applications
- Advanced Image Fusion Techniques
- Adversarial Robustness in Machine Learning
- Hydraulic and Pneumatic Systems
- Numerical methods in engineering
- Domain Adaptation and Few-Shot Learning
- Machine Fault Diagnosis Techniques
- Cellular and Composite Structures
Guangxi University
2025
Waseda University
2020-2024
Nanjing University of Aeronautics and Astronautics
1996-2021
Most methods based on the convolutional neural network show satisfying performance for hyperspectral image (HSI) classification. However, spatial dependence among different pixels is not well learned by CNNs. A recurrent (RNN) can effectively establish of nonadjacent and ensure that each feature activation in its output an at specific location concerning whole image, contrast to usual local context window recent limited conversion schemes RNN-based HSI classification cannot fully capture...
After existing ultra-deep vertical rotary tillers work in sugarcane stubble fields, the chopping performance is poor, and reason for this unknown. To solve this, paper develops a simulation model of tillage (UDVRT) field using FEM-SPH coupling method physical testing. The used to investigate process mechanism UDVRT cutter, identifying causes inadequate effectiveness. results show that, when comparing with test, magnitude variation cutter’s torque curves are relatively consistent, relative...
Strain information contains important dynamic characteristics of structure and could reflect the health status change inside structure. In a narrow space, pipeline strain measurement be challenge in engineering. This paper employs sensing beam to perform non-contact measurement. this study, frequency response function is used assess relation between measured on true Then, we proposed dual-channel device that can applied space measure strain. The tested an experiment platform, test results...
The number of international benchmarking competitions is steadily increasing in various fields machine learning (ML) research and practice. So far, however, little known about the common practice as well bottlenecks faced by community tackling questions posed. To shed light on status quo algorithm development specific field biomedical imaging analysis, we designed an survey that was issued to all participants challenges conducted conjunction with IEEE ISBI 2021 MICCAI conferences (80 total)....
Abstract Various deep neural network architectures (DNNs) maintain massive vital records in computer vision. While drawing attention worldwide, the design of overall structure lacks general guidance. Based on relationship between DNN and numerical differential equations, we performed a fair comparison residual with higher order perspectives. We show that widely used strategy, constantly stacking small (usually, 2–3 layers), could be easily improved, supported by solid theoretical knowledge...
Housing price is one of the most concerning issues to public worldwide. Studying spatial characteristics Shanghai’s housing prices and their explanatory factors great practical significance, for Shanghai largest city in China serves as national economic center a global financial hub. By crawling point interest (POI) data from Lianjia Real Estate Gaode Map past decade applying multiscale geographically-weighted regression (MGWR) model, this study deeply explores main influencing variables...
Abstract In the present paper, a damage gradient model combing concept with theory of critical distance (TCD) is established to estimate fatigue lives notched metallic structures under multiaxial random vibrations. Firstly, kind structure designed, and biaxial vibration tests are carried out different correlation coefficients phase differences between two axes. Then, evaluated utilizing proposed numerical simulations. Finally, validated by experiment results tests. The comparison demonstrate...
These days, imbalanced datasets, denoted throughout the paper by ID, (a dataset that contains some (usually two) classes where one considerably smaller number of samples than other(s)) emerge in many real world problems (like health care systems or disease diagnosis systems, anomaly detection, fraud stream based malware detection and so on) these datasets cause under-training minority class(es) over-training majority class(es), bias towards classification process application. Therefore, take...
Two time domain models for fatigue life prediction under multiaxial random vibrations are developed on the basis of critical plane approach. Firstly, stress power spectral density matrix each node at notch root test specimen is obtained by vibration analysis with finite element method, and time-histories generated from randomization Then, predicted based damage plane, where cumulative value greatest. The minimum all nodes considered as specimen. Finally, proposed validated 6061-T4 aluminum...
Abstract This paper aims to develop an innovative multiaxial fatigue criterion based on the stress invariant, in order perform damage analysis of engineering structures subjected random vibrations. First, change rule structural vibration lifetime under different load correlation conditions is systemically studied by experiments. Then, finite element such a structure, including modal and spectral analysis, implemented compute response power density (PSD) functions at critical fatigue‐prone...
Abstract The time domain critical plane‐based Carpinteri et al. criterion and the Theory of Critical Distance are here combined together to propose a novel procedure for vibration fatigue analysis circumferentially notched specimens subjected coupled multiaxial random environments. A metallic testing equipment is specially devised specimens; meanwhile, biaxial environment test implemented on in order investigate how both coherence phase shift between auto‐spectral density functions different...
Deep neural networks (DNNs) have extensively promoted data generation development; the quality of these generated content has achieved an impressive new level. Therefore, manipulated content, especially facial manipulation, is a growing concern for online information legitimacy. Most current deep learning-based methods depend on local features sampled by convolutional kernels and lack knowledge globally. To address problem, we propose dual-path pipeline using Neural Ordinary Differential...
Abstract Reformulation of a critical plane‐based multiaxial fatigue criterion named as Modified Wöhler Curve Method (MWCM) in the frequency domain is proposed; moreover, it combined with theory distance (TCD) to implement vibration lifetime prediction notched specimens subjected complex random loading. Firstly, frequency‐domain formulation maximum variance method (MVM) proposed for determining plane high computational efficiency. Then, reference curve defined through stress ratio...
Abstract For visual estimation of optical flow, which is crucial for various vision analyses, unsupervised learning by view synthesis has emerged as a promising alternative to supervised methods because the ground-truth flow not readily available in many cases. However, likely be unstable when pixel tracking lost via occlusion and motion blur, or correspondence impaired variations image content spatial structure over time. Recognizing that dynamic occlusions object usually exhibit smooth...
Vision Transformer (ViT) has been introduced into the computer vision (CV) field with its self-attention mechanism to capture global dependency. However, simply deploying ViT on a hyperspectral image (HSI) classification task can not get satisfying results because is spatial-only model, but rich spectral information exists in HSI. Moreover, most HSI classifiers integrate and spatial features cascaded flowchart, ignoring internal correlation between information. Furthermore, existing...
As frequency-varying sine excitations in rotating machines are always emerging during run-ups and shutdowns, the multi-input-multi-output (MIMO) swept-sine test is of utter significance product validation. At present, vibration tests mostly conducted with frequency-domain methods, where drive spectra generated updated by frequency response function (FRF), signals then sinusoid oscillators. In this paper, a time-domain approach using an inverse system method based on multistep prediction...
Vision Transformer (ViT) has recently been introduced into the computer vision (CV) field with its self-attention mechanism and gotten remarkable performance. However, simply applying ViT for hyperspectral image (HSI) classification is not applicable due to 1) a spatial-only model, but rich spectral information exists in HSI; 2) needs sufficient training samples, HSI suffers from limited samples; 3) does well learn local features; 4) multi-scale features are considered. Furthermore, methods...