A5101948335- Medical Research and Treatments
- Video Surveillance and Tracking Methods
- Face and Expression Recognition
- Advanced Computational Techniques and Applications
- Face recognition and analysis
- Traditional Chinese Medicine Studies
- Advanced Image and Video Retrieval Techniques
- Visual Attention and Saliency Detection
- Chaos-based Image/Signal Encryption
- Software Engineering Research
- Medical Image Segmentation Techniques
- Biometric Identification and Security
- Service-Oriented Architecture and Web Services
- Advanced Algorithms and Applications
- Advanced Decision-Making Techniques
- Traditional Chinese Medicine Analysis
- Impact of Light on Environment and Health
- Image Enhancement Techniques
- Advanced Sensor and Control Systems
- Software Reliability and Analysis Research
- Image Retrieval and Classification Techniques
- Advanced Neural Network Applications
- Infrared Target Detection Methodologies
- Acupuncture Treatment Research Studies
- Machine Learning in Healthcare
Chengdu Research Base of Giant Panda Breeding
2024
Capital Normal University
2023
Baidu (China)
2019-2023
Wenhua College
2022
Hebei Normal University
2006-2022
Wuhan College
2022
Xinjiang Production and Construction Corps
2022
Huazhong University of Science and Technology
2014-2021
Vision Technology (United States)
2019-2021
IM Systems (United States)
2020
The Visual Object Tracking challenge VOT2019 is the seventh annual tracker benchmarking activity organized by VOT initiative. Results of 81 trackers are presented; many state-of-the-art published at major computer vision conferences or in journals recent years. evaluation included standard and other popular methodologies for short-term tracking analysis as well methodology long-term analysis. was composed five challenges focusing on different domains: (i) VOTST2019 focused RGB, (ii)...
Porcine astrovirus (PoAstV) is widely distributed worldwide, and highly prevalent among piglets with or without diarrhea, existing as at least five distinct lineages (PoAstV1–PoAstV5) within the genus Mamastrovirus. However, our knowledge of diversity epidemiology PoAstV in China limited. In this study, fecal samples from 21/120 (17.5 %) domestic pigs, including 18/100 (18 diarrheic 3/20 (15 healthy 1/9 (11.1 wild boars tested Sichuan Province were positive for on reverse transcription–PCR....
We investigated the impact of early enteral nutrition (EEN) and parenteral (PN) on prealbumin (PA) high-sensitivity C-reactive protein (hs-CRP) in patients after gastric cancer surgery. Sixty-eight selected undergoing surgery were randomly divided into EEN (N = 34) PN groups. Body weight (BW), serum albumin (ALB), transferrin (TF), PA, hs-CRP, length hospital stay, cost postoperative nutritional support, incidence complications compared between On day 7, BW, TF, ALB, PA for both groups...
Model update lies at the heart of object tracking. Generally, model is formulated as an online learning problem where a target learned over training set. Our key innovation to \emph{formulate in meta-learning framework and learn algorithm itself using large numbers offline videos}, i.e., \emph{learning update}. The updater takes input set outputs updated model. As first attempt, we design based on recurrent neural networks (RNNs) demonstrate its application template-based tracker correlation...
A series of oleanolic acid derivatives were synthesized by diverse reactions, including the introduction conjugated alkadiene and epoxy ring moieties formed means photosensitized oxidation. Eosin Y was used as photosensitizer during this process. Next cytotoxicity products evaluated on HepG2.2.15 cells to determine appropriate treatment concentration for subsequent experiments. Most OA exhibited anti-HBV antigens secretion activity in cells. Among tested compounds, OA-4 (3.13 µg/mL) showed...
A new anticancer ligustrazine derivative, 3β-hydroxyolea-12-en-28-oic acid- 3,5,6-trimethylpyrazin-2-methylester (T-OA, C38H58O3N2), was previously reported. It synthesized via conjugating hepatoprotective and ingredients of traditional Chinese medicine. We found that T-OA exerted its activity by preventing the expression nuclear transcription factor NF-κB/p65 COX-2 in S180 mice. However, selective cytotoxicity on various kinds cell lines has not been studied sufficiently. In present study,...
Learning discriminative representation using large-scale face datasets in the wild is crucial for real-world applications, yet it remains challenging. The difficulties lie many aspects and this work focus on computing resource constraint long-tailed class distribution. Recently, classification-based learning with deep neural networks well-designed losses have demonstrated good recognition performance. However, memory cost linearly scales up to number of identities (classes) training set,...
Hepatic fibrosis is a naturally occurring wound-healing reaction, with an imbalance of extracellular matrix (ECM) during tissue repair response, which can further deteriorate to hepatocellular carcinoma without timely treatment. Inhibiting activated hepatic stellate cell (HSC) proliferation and inducing apoptosis are the main methods for treatment liver fibrosis. In our previous study, we found that TOA-glycine derivative (G-TOA) had exhibited more significant inhibitory activity against...
There has been no remarkable progress in the synthesis of sultones recent years. To facilitate more detailed studies this functional group, we found a new method to synthesize sulfonic acid lactone derivatives and finish its ring-closing reaction. A sultone derivative, (E)-ethyl 4-oxo-6-styryl-3,4-dihydro-1,2-oxathiine-5-carboxylate 2,2-dioxide (S-CA), was synthesized structurally identified by 1H-NMR, 13C-NMR, HMQC X-ray single crystal diffraction analysis. The rapid extended reaction...
Face detection is a challenging problem in the field of computer vision and key component machine intelligence. Viola Jones put forward face framework 2001 that can process images extremely rapidly while achieving high rates. This paper extends to handle rotated faces through simple improvement. The method rotate detected sub-windows by 30 degree clockwise counterclockwise within image plane. Experimental results show proposed algorithm has good rate reduces false positives regardless rotation faces.
The National Oceanic and Atmospheric Administration’s (NOAA) cloud-permitting high-resolution operational Hurricane Weather Research Forecasting (HWRF) model includes the sophisticated hybrid grid-point statistical interpolation (GSI) Ensemble Kalman Filter (EnKF) data assimilation (DA) system, which allows assimilating aircraft observations in tropical cyclone (TC) inner core regions. In HWRF DA flow-dependent background error covariance matrix is calculated from self-cycled 40-member...
Background subtraction is very important part of surveillance applications for successful segmentation objects from video sequences. The robust initial background extraction crucial in any subtraction. In this paper, we propose an algorithm to extract videos using dual frame differences and morphological processing. With the proposed algorithm, can be extracted accurately quickly. Experimental results various environmental sequences are provided demonstrate robustness, accuracy effectiveness...
Music feature extraction is widely used in music recommendation system.The recommended somewhat similar the form of melody.This phenomenon exhibits a repeating pattern that between set and one, which reveals has characteristics fractal dimension.In this paper, we selected time energy, frequency Mel-Frequency Cepstral Coefficient (MFCC) spectral envelope as features.These four features were integrated vectors calculated dimension by Hilbert transform.Compared with traditional content-based...
Fast appearance variations and the distractions of similar objects are two most challenging problems in visual object tracking. Unlike many existing trackers that focus on modeling only target, this work, we consider transient whole scene. The key insight is correspondence spatial layout scene consistent (i.e., global structure consistency) consecutive frames which helps to disambiguate target from distractors. Moreover, enables localize under fast variations. Specifically, propose an...
The purpose of attribute network representation learning is to learn the low-dimensional dense vector nodes by combining structure and information. current methods have insufficient interaction with when information, information cannot be well integrated. In this paper, we propose an method for dual-channel autoencoder. One channel structure, adopting multi-hop attention mechanism used capture node’s high-order neighborhood calculate weight; other node a low-pass Laplace filter designed...
Fast appearance variations and the distractions of similar objects are two most challenging problems in visual object tracking. Unlike many existing trackers that focus on modeling only target, this work, we consider \emph{transient whole scene}. The key insight is correspondence spatial layout scene consistent (i.e., global structure consistency) consecutive frames which helps to disambiguate target from distractors. Moreover, transient enables localize under fast variations. Specifically,...