A5102814159- Bioactive Natural Diterpenoids Research
- Biological Activity of Diterpenoids and Biflavonoids
- Phytochemistry and Biological Activities
- Medical Imaging Techniques and Applications
- Plant-derived Lignans Synthesis and Bioactivity
- Medical Image Segmentation Techniques
- Traditional and Medicinal Uses of Annonaceae
- Advanced MRI Techniques and Applications
- Biomedical Text Mining and Ontologies
- Plant-based Medicinal Research
- Image Retrieval and Classification Techniques
- Human Pose and Action Recognition
- Topic Modeling
- Anomaly Detection Techniques and Applications
- Remote-Sensing Image Classification
- Speech and Audio Processing
- Geochemistry and Geologic Mapping
- Currency Recognition and Detection
- Advanced Text Analysis Techniques
- Stock Market Forecasting Methods
- Speech Recognition and Synthesis
- Marine Sponges and Natural Products
- Handwritten Text Recognition Techniques
- Hand Gesture Recognition Systems
- Advanced Measurement and Detection Methods
Donghua University
2025
Hefei University of Technology
2009-2023
Chinese Academy of Sciences
2012-2019
Changchun University of Science and Technology
2019
Northwest University
2018-2019
Institute of Computing Technology
2019
Zhejiang Radio and Television University
2019
Wuhan University of Science and Technology
2014
Kunming Institute of Botany
2012-2013
Beijing University of Technology
2013
Machine learning has been increasingly prevailing all over the world, especially in computer vision field. This paper mainly focused on performance of MobileNetV2 model for image classification. To verify advanced better, this adopted MobileNetVl as control group and introduced an experiment identifying images a variety datasets extracted from TensorFlow. With T-SNE visualization tool, conclusion can be generated by comparing accuracy effectiveness these two models. The experimental results...
Drug-drug interactions (DDIs) may bring huge health risks and dangerous effects to a patient’s body when taking two or more drugs at the same time within certain period of time. Therefore, automatic extraction unknown DDIs has great potential for development pharmaceutical agents safety drug use. In this article, we propose novel recurrent hybrid convolutional neural network (RHCNN) DDI from biomedical literature. embedding layer, texts mentioning entities are represented as sequence...
When a patient takes two or more drugs within certain time, the efficacy of one drug may be influenced by other. This phenomenon is called drug-drug interactions (DDIs). DDIs are important and helpful information for both medical staff patients to make sure that co-administrated at same time have positive effect on therapy patients. Many approaches been applied into interaction extraction tasks such as support vector machine (SVM), recurrent neural network (RNN) long short-term memory (LSTM)...
Eleven new dibenzocyclooctadiene lignans, polysperlignans A-K (1-11), and eight known analogues (12-19) were isolated from the stems of Kadsura polysperma. Their structures absolute configurations established using a combination MS, NMR, CD, single-crystal X-ray diffraction techniques. Selected compounds evaluated for activity against β-amyloid- or hydrogen peroxide-induced neurotoxicity on PC12 cells, 1, 2, 4, 5, 13, 16 showed statistically significant neuroprotective effects in these vitro assays.
Scopariusins A–C (1–3), three novel rearranged ent-halimanoids with a bicycle[5.4.0]undecane ring system, two new normal (4 and 5), ent-clerodanoid (6) were isolated from Isodon scoparius. Moreover, biomimetic transformation the to ent-halimane diterpenoids was successfully accomplished, which not only validated biogenetic hypothesis in this plant but also confirmed absolute configurations of 1 5.
Isorosthin A (1), the first 20-nor-enmein-type diterpenoid, and 15 new ent-kauranoids, isorosthins B-P (2-16), along with 22 known analogues were isolated from aerial parts of Isodon rosthornii. The structures 1-16 elucidated by means spectroscopic analysis. relative configuration 2 absolute 3 determined single-crystal X-ray diffraction. Cytotoxicity evaluation against five human tumor lines showed inhibitory effects several compounds tested. Furthermore, 12 isolates exhibited activity...
A new ent-atisanoid (1) and three ent-kauranoids (2-4) belonging to different types, along with four known compounds were isolated from Isodon rosthornii. Their structures established by means of extensive spectroscopic analysis. The absolute configuration 1 was further determined X-ray diffraction. Compounds 5 are the first example atisane-type diterpenoid this plant.
Human-computer interaction is a hot topic in artificial intelligence. Artificial navigation an interesting application of human-computer interaction, which control the action target device by speech or gestures information. The main virtue that it can within distance without any remote device. This technology be used areas robot navigation, vehicle industrial site and virtual reality. paper proposed algorithm for combining gesture recognition with recognition. Firstly, use nine instructions...
In classification of fashion article images based on e-commerce image recommendation system, the accuracy and computation time cannot meet actual requirements. Herein, for first to our knowledge, we present two diverse recognition approaches called random-forest method genetic algorithm (GA-RF) Visual Geometry Group-Image Enhancement (VGG-IE) solve problem. GA-RF, number segmentation times decision trees are key factors affecting results. Improved is introduced into parameter optimization...
In recent years Transfer Learning has become an important research issue and it is employed to solve the lack of data problem in remote sensing field. As one kind most popular generative models, Generative Adversarial Networks (GANs) can generate hierarchy representations realistic textures objects. Now GANs have successful usual way synthesize images. this paper we train different images on UC Merced Land Use Dataset aircraft carrier dataset compiled by ourselves. Our are better than ones...
We evaluate the performance and accuracy of different motion correction algorithms used in nuclear cardiac imaging. Three automated tracking techniques are diverging square (DS); circles (DC); self-adaptive masking methods (SM). The observation process over images acquired (specifically for this study with introduced patient motion) laboratory showed that DC SM perform better when compared DS algorithm. It was shown method gains a good tolerance low signal to noise ratio (SNR) is also robust...
We introduce a new self-adaptive masking method for shape detection of low signal to noise ratio (SNR) images improve the tracking capabilities motion correction in nuclear cardiac imaging. The is developed using two-dimensional fast Fourier transform, ideal filtering frequency domain, recursive thresholding, and region recognition. This independent correlation between context planar has good tolerance SNR images. Also it robust under circumstances significant abrupt object.
Nowadays many E-commerce corporations need to implement effective content management for controlling the data resources of their Websites. In this paper, a role-authorized and MVC-based method is proposed. Firstly, presentation mechanism designed based on MVC model. Secondly, system users separated into three types, including administrator, manager visitor, then model designed. The proposed can bring good performance corporations.
In1 recent years, deep neural network has been widely used in image recognition, speech processing and natural language processsing, which achieved good results. In order to solve the problem of fast screening tobacco-related information from Internet, paper proposes a informative algorithm based on convolutional network. Firstly, term frequency-inverse document frequency (Tf-idf) Simplified (Sdcnn) are extract text features information, then images texts effectively fused. Experiments show...
The change of field view (FOV) which is caused by the image defocus laser communication terminal introduced in process assembly will have a great interference to link energy budget. In problem, method measuring FOV proposed calculate amount receiving system. On premise sensitivity detector satisfied received energy, relationship among and are derived theoretical analysis calculation. With distance 1km, 0.27mm calculated under 0.965mrad measured experimental test. maximum from 2km 4km can be...