A5057996314- Cellular and Composite Structures
- Acoustic Wave Phenomena Research
- Noise Effects and Management
- Remote-Sensing Image Classification
- Additive Manufacturing and 3D Printing Technologies
- Aluminum Alloys Composites Properties
- Composite Structure Analysis and Optimization
- Bone Tissue Engineering Materials
- Advanced Image Fusion Techniques
- Advanced Neural Network Applications
- Magnesium Alloys: Properties and Applications
- Stock Market Forecasting Methods
- Electric Motor Design and Analysis
- Advanced Text Analysis Techniques
- Superconducting Materials and Applications
- Traditional Chinese Medicine Analysis
- Vehicle Noise and Vibration Control
- Additive Manufacturing Materials and Processes
- Barrier Structure and Function Studies
- Hearing Loss and Rehabilitation
- Vehicle License Plate Recognition
- Domain Adaptation and Few-Shot Learning
- Ichthyology and Marine Biology
- Titanium Alloys Microstructure and Properties
- Shape Memory Alloy Transformations
Shandong Academy of Sciences
2023-2024
Qilu University of Technology
2023-2024
Shandong University
2023-2024
Harbin Engineering University
2019-2023
Anhui Polytechnic University
2022-2023
Yueyang Hospital
2022
Shanghai University of Traditional Chinese Medicine
2022
Tsinghua University
2021
The fusion classification of hyperspectral image (HSI) and light detection ranging (LiDAR) data has gained widespread attention because its ability to obtain more comprehensive spatial spectral information. However, the heterogeneous gap between HSI LiDAR also adversely affects performance. Despite excellent performance traditional multimodal models, language information containing much linguistic priori knowledge enrich visual representations needs be addressed. Therefore, we design a...
Buzhong Yiqi decoction (BZYQD) has been developed for preventing or reducing the recurrence of ischemic stroke a long time in China. However, mechanism action BZYQD is not completely understood. Our research aims to determine whether by regulating gut microbiota using 16SR RNA and fecal transplantation. In cerebral ischemia mouse model, results showed that prophylactic administration could reduce brain infarct volume improve neurological function behavior. The regulate intestinal increase...
With the development of underwater vehicle technology, application target detection ranges from ocean search, rescue and to fishing for marine products. At this stage, main problem seafood is slow processing time caused by a large number parameter calculations in deep learning networks. This paper proposes lightweight network based on YOLOv5s. First, GhostBottleneck used reduce parameters. Second, convolution block attention mechanism added last layer backbone increase weight valuable...
Abstract In this article, 316L stainless steel hollow spheres (316L HS) and polyurethane (PU) resins were used to prepare a novel metallic‐hollow‐sphere/polyurethane (MHSP) acoustic composite by casting. The experimental results of revealed that compared monolithic PU, the sound transmission loss composites increased maximum absorption peak shifted lower frequency with change impedance resonance frequency. When surface metallic sphere in MHSP was modified silane coupling agent, insulation...
Abstract In this investigation, a variety of polyurethane composites with different syntactic hollow spheres were prepared by casting method, using 316 L stainless steel and steel–Al 2 O 3 double‐layer as reinforcements. The sound insulation absorption properties tested impedance tube. effects the on acoustic analyzed compared resin polyurethane/solid composites. experimental results reveal that transmission loss can be improved addition spheres, range frequency tends to move lower...
In practical hyperspectral image (HSI) classification tasks, we often encounter the problems of few-shot and domain misalignment between source domains target domains. To solve this paradigm, a meta-learning method learning (FSL) is usually used. However, most existing FSL methods address problem for alignment neglect exploration semantic relationships objects across paper, propose cross-domain spectral transformer (SSTFSL), which can fully extract features detail HSI task. Specifically,...
Additive manufacturing (AM) usually leads to the epitaxial growth of columnar grains in deposition layer along construction direction, typically resulting decrease mechanical properties. In this study, A517Q alloy steel was fabricated by laser and wire additive (LWAM) assisted with ultrasonic vibration (UV) treatment. The microstructure characteristics under different UV affected were investigated OM, XRD, SEM, EBSD. Compared sample without UV, percentage equiaxed samples dual increased 90%....
Abstract As a noval functional material designed and prepared artificially, one‐dimensional phononic crystal has great prospect in the field of noise reduction. In this study, novel metamaterial based on metallic hollow spheres (MHS)/polymer matrix laminated composites was prepared. The result impedance tube test shows that composite excellent sound insulation performance. incident direction waves little effect performance composites. average MHSP/SR‐LC (MHSP side) 44.4% higher than SR...
Object detection is extensively used in sectors as intelligent transportation and monitoring, however the current general single shot multibox detector (SSD) has drawback of being unable to detect small objects. This study proposes an improved SSD enhance accuracy objects, dubbed RFP-SSD. It adopts Resnet50 backbone network fuses attention mechanism boost shallow feature extraction stage. Further, a recursive pyramid (RFP) structure applied facilitate multiscale fusion among maps. The...
Deep learning models have been widely used in hyperspectral image (HSI) classification tasks. However, the convolutional neural network-based typically suffer from low utilization of spectral sequence information and poor processing power detail information. In contrast, transformers are able to learn spectrum, but they either focus only on local feature extraction or global extraction. To address this issue, we propose transformer (LGST), which consists a swin block, fusion (LGF) structure....
Over the past decades, industry classification has been focus of research due to its potential be used in rotation trading strategies, bringing return beat market. However, on Chinese A-shares is fairly scarce, and many them rely heavily human interference, leading difficulties replicating experiments classify firms. This paper proposed a text-based method tackle this issue, which can automatically To specific, we employ hierarchical clustering firm annual reports combined with financial...