A5041829815- Carbon Dioxide Capture Technologies
- Phase Equilibria and Thermodynamics
- Membrane Separation and Gas Transport
- Adversarial Robustness in Machine Learning
- Advanced SAR Imaging Techniques
- Bacillus and Francisella bacterial research
- Mesoporous Materials and Catalysis
- Multicomponent Synthesis of Heterocycles
- Zeolite Catalysis and Synthesis
- Microbial bioremediation and biosurfactants
- Image Processing Techniques and Applications
- Carbon dioxide utilization in catalysis
- Remote-Sensing Image Classification
- Domain Adaptation and Few-Shot Learning
- Covalent Organic Framework Applications
- Synthesis and Biological Evaluation
- High-Velocity Impact and Material Behavior
- CO2 Sequestration and Geologic Interactions
- Phytochemistry and Bioactivity Studies
- Natural product bioactivities and synthesis
- Synthetic Aperture Radar (SAR) Applications and Techniques
- Advanced Image Processing Techniques
- Inorganic and Organometallic Chemistry
- Digital Media Forensic Detection
- Crystallization and Solubility Studies
National University of Defense Technology
2021-2024
Jiangnan University
2018-2022
State Key Laboratory of Clean Energy Utilization
2021-2022
Zhejiang Energy Research Institute
2021-2022
Zhejiang University
2017-2022
CSIRO Manufacturing
2021-2022
Guangxi University for Nationalities
2022
First People's Hospital of Nanning
2022
Shanghai University
2016-2019
Northeast Normal University
2016-2017
Abstract Organic lyotropic liquid‐crystal (LLC) assemblies mimic molecular sieves in their nanoporous structures and ability to incorporate catalytic functional groups. This article focuses on recent advances made by our research group incorporating new properties into polymerizable LLC studying the molecular‐transport of crosslinked networks.
Streptococcus mutans has been reported as a primary cariogenic pathogen associated with dental caries. The bacteria can produce glucosyltransferases (Gtfs) to synthesize extracellular polysaccharides (EPSs) that are known virulence factors for adherence and formation of biofilms. Therefore, an ideal inhibitor caries is one inhibit planktonic growth prevent biofilm formation. Bergenia crassifolia (L.), widely used folk medicine tea beverage, have variety bioactivities. present study aimed...
The catalytic reactivity and selectivity of the first example a nanostructured solid acid resin (1) are described. This new type catalyst is formed by self-assembly copolymerization two acidic lyotropic liquid crystals (LLCs), affording columnar hexagonal polymer network with monodisperse nanochannels lined sulfonic groups. performance this material as heterogeneous was compared against that commercially available, amorphous resins: Amberlyst-15 Nafion NR50. Using acid-catalyzed...
How to simultaneously realize retrieving oil and recycling surfactant in the remediation of leaked oil-polluted soil by means surfactant-enhanced washing is still a significant challenge. Here, we reported for first time novel CO2-switchable anionic surfactant, 11-dimethylamino-undecyl sulfate sodium salt (DUSNa), retrieve recycle process sand DUSNa-enhanced washing. Because tertiary amine group had been incorporated into traditional alkyl form DUSNa, DUSNa was readily converted its inactive...
SAR-optical images from different sensors can provide consistent information for scene classification. However, the utilization of unlabeled in deep learning-based remote sensing image interpretation remains an open issue. In recent years, contrastive self-supervised learning (CSSL) methods have shown great potential obtaining meaningful feature representations massive amounts data. This paper investigates effectiveness CSSL-based pretraining models remote-sensing Firstly, we analyze...
Synthetic aperture radar (SAR) can perform observations at all times and has been widely used in the military field. Deep neural network (DNN)-based SAR target recognition models have achieved great success recent years. Yet, adversarial robustness of these received far less academic attention remote sensing community. In this article, we first present a comprehensive evaluation framework for DNN-based recognition. Both data-oriented metrics model-oriented to fully assess performance under...
Several taper-shaped, Brønsted acidic amphiphiles containing an amide linkage near the headgroup were synthesized with goal of generating acidic, inverted hexagonal (HII) lyotropic liquid crystal (LLC) phases that can be stabilized by polymerization. These polymerizable reacting acyl chloride 3,4,5-tris(11'-acryloxyundecyloxy)benzoic acid several amino acids (l-alanine, β-alanine, l-phenylalanine) and 2-aminoethanesulfonic acid. Of these derivatives, only l-alanine derivative (1) was found...
Abstract The industrial application of emerging water‐lean solvents to CO 2 capture from flue gas is challenged by their high viscosity. In this work, we report a novel solvent which possesses lower viscosity and higher cyclic capacity than other reported in the literature. new consists N, N ‐dimethyl‐1, 2‐ethanediamine (DMEDA), physical cosolvent ‐methyl‐2‐pyrrolidone (NMP) up 15% water (named ENH). We evaluated effect composition on viscosity, regeneration energy ENH compared it with...
Deep neural networks (DNNs) have been widely utilized in automatic visual navigation and recognition on modern unmanned aerial vehicles (UAVs), achieving state-of-the-art performances. However, DNN-based systems UAVs show serious vulnerability to adversarial camouflage patterns targets well-designed imperceptible perturbations real-time images, which poses a threat safety-related applications. Considering scenario UAV is suffering from attack, this paper, we investigate construct two...
Abstract A facile multi‐component process for the synthesis of perfluoroalkylated quinolizine derivatives was achieved using various arylidenemalononitriles, pyridine, and methyl perfluoroalk‐2‐ynoates as starting materials. Moderate yields were obtained under mild condition. The structures characterized by means 1 H NMR, 13 C 19 F IR, LRMS HRMS. Furthermore, reaction mechanism proposed.
Adversarial training is an effective method to enhance adversarial robustness for deep neural networks. However, it qequires large amounts of labeled data, which are often difficult acquire. Recent research has shown that self-supervised learning can help improve model performance and uncertainty using unlabeled data. In this paper, we introduce a new framework learn robust pretrained remote sensing scene classification. The proposed exploits the advantage dual network structure, requires...
Significant reduction of the water content traditional absorbents, increasing organic character absorbent molecules, and substitution with a non-aqueous diluent are increasingly attracting interest as means to improve performance. From our previous work, novel diamine absorbents N,N-dimethyl-1,3-propanediamine (DMPDA) N,N-dimethyl-1,2-ethanediamine (DMEDA), also utilizing N-methyl-2-pyrrolidone (NMP) reduce absorbent, were demonstrated produce an blend significantly lower overall energy...
Robustness, both to accident and malevolent perturbations, is a crucial determinant of the successful deployment deep convolutional neural network based SAR ATR systems in various security-sensitive applications. This paper performs detailed adversarial robustness evaluation models across two public available target recognition datasets. For each model, seven different ranging from gradient optimization self-supervised feature distortion, are generated for testing image. Besides average...
Super-resolution reconstruction technology is an important research topic in many fields such as image processing and computer vision. This can be used widely for security monitoring, old reconstruction, compression transmission other fields. In this paper, super-resolution performed on a low-resolution of four times magnification. We propose the dense convolutional networks generator instead residual networks, set perceptual loss optimization goal. use VGG network feature map function Mean...
Synthetic aperture radar (SAR) images often suffer from sample missing problems, which are a consequence of the high cost imaging. However, deep learning methods heavily rely on large-scale, high-quality labelled data, while traditional feature-based classification requires manual classifier design. To resolve above limitations, we proposed SAR target recognition method based feature fusion and spiking neural network (FF-SNN). Initially, extract features fuse these with different strategies....