A5100413039- Topic Modeling
- Text and Document Classification Technologies
- Video Surveillance and Tracking Methods
- Robotics and Sensor-Based Localization
- High-pressure geophysics and materials
- Advanced Vision and Imaging
- Robot Manipulation and Learning
- Human Pose and Action Recognition
- Advanced Image and Video Retrieval Techniques
- Sentiment Analysis and Opinion Mining
- Advanced Text Analysis Techniques
- Radiomics and Machine Learning in Medical Imaging
- Data Quality and Management
- Face and Expression Recognition
- 3D Surveying and Cultural Heritage
- Image Retrieval and Classification Techniques
- Advanced Graph Neural Networks
- Neuroscience and Neuropharmacology Research
- Neural Networks and Applications
- Lipid metabolism and disorders
- COVID-19 diagnosis using AI
- Lung Cancer Diagnosis and Treatment
- Diamond and Carbon-based Materials Research
- Brain Tumor Detection and Classification
- Obstructive Sleep Apnea Research
Beijing University of Technology
2019-2024
Nanjing Tech University
2021-2024
Jinan University
2023-2024
Second Affiliated Hospital of Nanchang University
2022-2024
Nanchang University
2022-2024
Peking University
2001-2024
Peking University People's Hospital
2024
Nagoya University
2023
Nagoya University Hospital
2023
Shandong First Medical University
2022
The Vision Meets Drone (VisDrone2019) Single Object Tracking challenge is the second annual research activity focusing on evaluating single-object tracking algorithms drones, held in conjunction with International Conference Computer (ICCV 2019). VisDrone-SOT2019 Challenge goes beyond its VisDrone-SOT2018 predecessor by introducing 25 more challenging sequences for long-term tracking. We evaluate and discuss results of 22 participating 19 state-of-the-art trackers collected dataset. are...
Abstract Utilizing the magnetic charge theory, analytical expressions of shape demagnetization tensor spatially profile a cuboid are derived and validated. The can accurately calculate factors in three orthogonal directions for given point rectangular body. is function positions dimension ratios independent material permeability. significant convenience analyzing designing optimizing cuboid.
Abstract Lung cancer is mainly caused by malignant lung nodules. Early detection and diagnosis of nodules can diagnose the disease in time significantly improve survival rate patients. With rapid development deep learning networks field medical aid diagnosis, many have been applied to nodule detection. Statistical distribution shows that most radii are too small be well detected. Therefore, 3D feature pyramid network (FPN) for single‐stage pulmonary proposed solve this problem combining...
Schizophrenia (SCZ) is a severe psychiatric disorder characterized by positive symptoms, negative and cognitive deficits. Current antipsychotic treatment in SCZ improves symptoms but has major side effects little impact on impairment. The pathoetiology of remains unclear, known to involve small GTPase signaling. Rho kinase, an effector Rho, highly expressed the brain plays role neurite elongation neuronal architecture. This study used touchscreen-based visual discrimination (VD) task...
In traditional Chinese medicine, Qi-zhi-wei-tong granule (QZWT) significantly reduced the major gastrointestinal and psychological symptoms of functional dyspepsia. Aim study: We aimed to explore therapeutic effect QZWT treated chronic non-atrophic gastritis (CNAG) elucidate its potential mechanism. The composition was analysed by UPLC-Q/TOF-MS. CNAG mice model established restraint stress (CRS) in combination with iodoacetamide (IAA). Morphological staining utilized reveal impact on stomach...
This paper proposes an efficient robot assembly skill learning framework based on only a few demonstrations. The process consists of two phases, e.g., the pre-training phase and self-learning phase. In phase, networks are initialized from demonstration data. A novel data augmentation model state transition is designed, which requires several human demonstrations to learn parameters reduces cost collection in pre-training. pre-trained further optimized by hybrid exploration strategy...
Abstract: One of the most pressing emerging issues in bacterial resistance is multidrug-resistant Klebsiella pneumoniae . The treatment K. infections often problematic because lack available therapeutic options, leading to negative effects on morbidity, mortality, and healthcare-associated costs. Carrimycin a macrolide antibiotic with good antibacterial effects. In this study, we report patient diagnosed infection who was treated carrimycin. presented cough, expectoration, dyspnea, severe...
Abstract Two new secolignans, peperomins G and H ( 1 2 , resp.), were isolated from the whole plant of Peperomia dindygulensis together with five known peperomin A 3 ), E 4 B 5 2,3‐ trans ‐2‐methyl‐3‐{(3‐hydroxy‐4,5‐dimethoxyphenyl)[5‐methoxy‐3,4‐(methylenedioxy)phenyl]methyl}butyrolactone 6 cis ‐2‐(hydroxymethyl)‐3‐{bis[5‐methoxy‐3,4‐(methylenedioxy)phenyl]methyl}butyrolactone 7 ). Their structures configurations elucidated by spectroscopic methods including 2D‐NMR techniques....
Early life stress correlates with a higher prevalence of neurological disorders, including autism, attention-deficit/hyperactivity disorder, schizophrenia, depression, and Parkinson's disease. These conditions, primarily involving abnormal development damage the dopaminergic system, pose significant public health challenges. Microglia, as primary immune cells in brain, are crucial regulating neuronal circuit survival. From embryonic stage to adulthood, microglia exhibit stage-specific gene...
Purpose The positioning and grasping of large-size objects have always had problems low accuracy, slow speed high application cost compared with ordinary small parts tasks. This paper aims to propose implement a binocular vision-guided system for object industrial robot. Design/methodology/approach To guide the robot grasp position pose this study measures by extracting reconstructing three non-collinear feature points on it. improve precision robustness measuring, coarse-to-fine strategy is...
Recently, Graph Convolutional Neural Network (GCN) is widely used in text classification tasks. And it has been effectively to accomplish tasks that are thought have a rich relational structure. However, due the sparse adjacency matrix constructed by GCN, GCN cannot make full use of context-dependent information classification, and not good at capturing local information. The Bidirectional Encoder Representation from Transformers (BERT) ability capture contextual sentences or documents, but...
Filoviruses such as Ebola virus (EBOV), Marburg (MARV), and Sudan (SUDV) cause deadly viral hemorrhagic fever in humans, with high case-fatality rates; however, no licensed therapeutic agent or vaccine has been clinically approved to treat prevent infection.T-705 (favipiravir) is a novel antiviral drug that for the treatment of influenza Japan.T-705 exhibits broad-spectrum activity against different viruses, including MARV EBOV, here, we are first report vitro vivo T-705 SUDV.T-705 reduced...
The high pressure phase transition at zero temperature and the phonon-dispersion relations temperatue pressue of BeO have been studied by a first-principles method. results show that from wurtzite structure (B4) to cubic sodium chloride (B1) happens about 122 GPa zinc blende (B3) is meta-stable pressure. B1, B3 B4 are investigated frozen phonon calculations B1 an unstable phases two very simliar structure, but they still distinguishable each other their relations. Finally, diagrams studied.
The shock-induced phase transformation of nanocrystalline iron with different grain sizes is investigated by using molecular dynamic simulations. critical shock stress for (from body-cubic centered α into hexagonal-close packed ε phase) irons about 15 GPa. Under compression, the first experience elastic deformation, then plastic deformation purely caused boundaries, after that nucleated mostly at and finally nucleation areas expanding entire samples. These processes can be reflected profile...
Text classification is a fundamental task in natural language processing. This widely concerned and applied. However, previous methods mainly use traditional static word embedding, but embedding could not deal with the problem of polysemy. For this reason, we propose to utilize contextualized BERT effectively encode input sequence then temporal convolutional module which simply computes 1-D convolution extract high-level features, finally, max-pooling layer retains most critical features for...