A5082621728- Advanced Neural Network Applications
- Advanced oxidation water treatment
- Thermal Radiation and Cooling Technologies
- TiO2 Photocatalysis and Solar Cells
- Brain Tumor Detection and Classification
- Domain Adaptation and Few-Shot Learning
- Advanced Photocatalysis Techniques
- Robotics and Sensor-Based Localization
- Advanced Image and Video Retrieval Techniques
- Concrete and Cement Materials Research
- Advanced Sensor and Energy Harvesting Materials
- Covalent Organic Framework Applications
- Microwave Engineering and Waveguides
- Structural Engineering and Vibration Analysis
- Perovskite Materials and Applications
- Enhanced Oil Recovery Techniques
- Extraction and Separation Processes
- Optical properties and cooling technologies in crystalline materials
- Industrial Automation and Control Systems
- Recycled Aggregate Concrete Performance
- Luminescence Properties of Advanced Materials
- Infrared Target Detection Methodologies
- Hydraulic Fracturing and Reservoir Analysis
- Textile materials and evaluations
- Building Energy and Comfort Optimization
Shanghai Institute of Technology
2021-2024
Donghua University
2023
Shanghai Shipbuilding Technology Research Institute
2010-2021
Yangtze University
2018
National Laboratory of Solid State Microstructures
2016
Waveguide superlattices, a special type of waveguide arrays, can be designed to achieve very low cross talk at submicrometer/subwavelength pitches. The theoretical framework and design rationales for such superlattices will presented in depth. sidewall roughness help deter the coherent coupling between identical waveguides nearby supercells, but it also induces random fluctuation transmission. Statistical behavior transmission due superlattice is systematically treated. Complex...
Temperature regulating fibers (TRFs) with high enthalpy and form stability are the key factors for thermal management. However, enthalpies of most TRFs not high, preparation methods still at laboratory scale. It remains a great challenge to use industrial spinning equipment achieve continuous processing excellent mechanical properties. Here, polyamide 6 (PA6) based sheath-core structure were prepared by bicomponent melt-spinning. The TRF (TRFsc) composed PA6 as sheath functional core, which...
Lightweight, high‐performance networks are important in vision perception systems. Recent research on convolutional neural has shown that attention mechanisms can significantly improve the network performance. However, existing approaches either ignore significance of using both types (channel and space) simultaneously or increase model complexity. In this study, we propose adaptive module (AAM), which is a truly lightweight yet effective comprises channel spatial submodules to balance...
The surface modification of pore throat by adsorption surfactants is thought to have a positive effect on water flooding in low and ultralow permeability reservoirs. In this paper, Gemini cationic surfactants, containing 12 16 carbon alkyl chains(ethanediyl-1,2-bis(dimethyl dodecyl ammonium bromide) ethanediyl-1,2-bis(dimethyl cetyl bromide), referred as GC12 GC16) hexadecyl trimethyl bromide (CTAB) were used modifying agents investigate the effects surfactant concentration, time temperature...
For unmanned aerial vehicle (UAV), object detection at different scales is an important component for the visual recognition. Recent advances in convolutional neural networks (CNNs) have demonstrated that attention mechanism remarkably enhances multiscale representation of CNNs. However, most existing feature methods simply employ several blocks to adaptively recalibrate response, which overlooks context information a level. To solve this problem, filtering network (MFFNet) proposed paper...
Recent research has shown that attention mechanisms can help convolutional networks train and infer more efficiently accurately. However, the current mechanism mainly focuses on relationship between global features ignores local features. A Feature Filtering Module (FFM) for neural is proposed in this paper. The FFM module uses both spatial channel dimensions fuses feature maps of two branches into a 3D map to effective information flow efficiently. Extensive tests CIFAR-100, MS COCO show...
Recent research has demonstrated that the multiscale representation of CNNs is considerably improved by attention mechanism. However, majority current feature techniques only employ a small number blocks in mechanism, ignoring multiscale-level contextual data. This suggests fused filtering module as solution to this issue (MFFFM). It enables branch-specific feature-selective learning These branches various sizes null convolutions, and they further spatial injection correlation channel...
Abstract In this paper, Fe 2 O3 power was used to replace the scrap steel in traditional process, as medium for removing free acid waste solution of cold-rolled strip pickling. The removal effect and silicon system with Fe2O3 addition crap were compared. results indicated that obtained a better effect. Subsequently, optimum dosage, reaction temperature time determined by orthogonal experiment. flocculants polyacrylamide (PAM), PAC, PAM+PAC MgCl2 desiliconization explored.