A5103093819- Machine Learning and ELM
- Image Enhancement Techniques
- Industrial Vision Systems and Defect Detection
- Nanofluid Flow and Heat Transfer
- Color Science and Applications
- Fluid Dynamics and Turbulent Flows
- Particle Dynamics in Fluid Flows
- Heat Transfer Mechanisms
- Advanced Vision and Imaging
- Computational Fluid Dynamics and Aerodynamics
- Microfluidic and Bio-sensing Technologies
- Heat Transfer and Optimization
- Face and Expression Recognition
- Video Surveillance and Tracking Methods
- Lattice Boltzmann Simulation Studies
- Textile materials and evaluations
- Fluid Dynamics and Heat Transfer
- Thermography and Photoacoustic Techniques
- Combustion and flame dynamics
- Solar Thermal and Photovoltaic Systems
- Gas Dynamics and Kinetic Theory
- Image Processing Techniques and Applications
- Modeling and Simulation Systems
- Water Quality Monitoring Technologies
- Advanced Memory and Neural Computing
Hangzhou Dianzi University
2015-2024
Zhejiang University
2022
Zhejiang Sci-Tech University
2005-2021
Zhejiang University of Science and Technology
2010-2016
Wuhan University of Science and Technology
2015
To improve efficiency and classification accuracy, overcome the issue of poor generalization performance traditional fabric defect methods, we present a new algorithm with evolving Inception v3 by improved L 2,1 -norm regularized extreme learning machine. Herein, first, features images are extracted using v3, which reduces amount computation improves accuracy feature extraction. Second, an L2,1-norm regularization machine based on multi-verse optimizer–Henry gas solubility optimization,...
It is known the thermal conductivity can be elevated considerably by adding nanoparticles in base fluid. But mechanisms remain unclear so far. One of most important to elevate aggregation nanoparticles. Molecular dynamics (MD) simulation normally employed model behavior very small system due great computational workload. In present work, a new hybrid method multi-particle collision and molecular (MPCD-MD) proposed evaluate morphology Cu-H2O nanofluid for reducing Such competent simulating...
To mitigate the problem of low classification accuracy in solid color printing and dyeing, a difference model based on differential evolution (DE) improved whale optimization algorithm (WOA) for extreme learning machine (ELM) optimization, named DE–WOA–ELM, was developed this study. Considering that initial population WOA has significant influence solution speed quality, DE used to generate more suitable by avoiding local optima, thereby improving performance. The method an excellent global...
With the more and amelioration of our quality life, needs for clothing have altered from having clothes to wearing good-looking, among which wrinkle resistance fabric owns a giant effect on beauty clothing. Nowadays, artificial subjective evaluation is mainly used evaluate grade garment fabrics in textile industry. This method shortcoming poor accuracy, being time-consuming objectivity. For solving this problem, it very important put forward an objective model grade. In paper, we proposed...
To ameliorate the precision of clothing image classification, we proposed a classification method via DenseNet201 network based on transfer learning and optimized regularized random vector functional link (RVFL). First, formula extracts weight's parameters about that is pre-trained ImageNet dataset for learning, thereby obtaining an incipient network,after trim this model parameters. The modified utilized to pick up features output by DenseNet201's global average pooling layer. Second,...
This article presents an intelligent algorithm based on extreme learning machine and sequential mutation genetic to determine the inverse kinematics solutions of a robotic manipulator with six degrees freedom. is developed minimize computational time without compromising accuracy end effector. In proposed algorithm, preliminary solution first computed by then optimized improved mutation. Extreme randomly initializes weights input layer biases hidden layer, which greatly improves training...
This study proposes an ensemble differential evolution online sequential extreme learning machine (DE-OSELM) for textile image illumination correction based on the rotation forest framework. The DE-OSELM solves inaccuracy and long training time problems associated with traditional algorithms. First, Grey–Edge framework is used to extract low-dimensional efficient features as (OSELM) input vectors improve speed of OSELM. Since weight hidden-layer bias OSELMs are randomly obtained, OSELM...
To improve the accuracy and efficiency of objective evaluation fabric wrinkle model, R18-COTSA-RVFL model is proposed in this paper. The rating based on combination ResNet18 (R18) enhanced random vector functional-link network, which network replaces Softmax layer ResNet18. This uses to extract features from wrinkled images. In addition, a improved by tunicate swarm algorithm was for wrinkle-level classification. First, we used chaotic-logistic maps opposition-based learning optimize initial...
To improve accuracy in clothing image recognition, this paper proposes a classification method based on parallel convolutional neural network (PCNN) combined with an optimized random vector functional link (RVFL). The uses the PCNN model to extract features of images. Then, structure-intensive and dual-channel (i.e., PCNN) is used solve problems traditional networks (e.g., limited data prone overfitting). Each layer followed by batch normalization layer, leaky rectified linear unit...