A5103260862- Advanced Neural Network Applications
- Vehicle License Plate Recognition
- Image Processing and 3D Reconstruction
- Handwritten Text Recognition Techniques
- Respiratory viral infections research
- Video Surveillance and Tracking Methods
- Blind Source Separation Techniques
- Planarian Biology and Electrostimulation
- Nanoparticles: synthesis and applications
- Artificial Intelligence in Healthcare
- Tunneling and Rock Mechanics
- melanin and skin pigmentation
- Sentiment Analysis and Opinion Mining
- Text and Document Classification Technologies
- Spam and Phishing Detection
- Microplastics and Plastic Pollution
- Rough Sets and Fuzzy Logic
- Biological and pharmacological studies of plants
- Cutaneous Melanoma Detection and Management
- Water Quality Monitoring Technologies
- Medical Image Segmentation Techniques
- Image Enhancement Techniques
- Advanced Adaptive Filtering Techniques
- Speech and Audio Processing
- Remote Sensing and LiDAR Applications
Southwest Minzu University
2021-2025
State Ethnic Affairs Commission
2021-2024
Beijing Center for Disease Prevention and Control
2018-2023
Malicious websites present a substantial threat to the security and privacy of individuals using internet. Traditional approaches for identifying these malicious sites have struggled keep pace with evolving attack strategies. In recent years, language models emerged as potential solution effectively detecting categorizing websites. This study introduces novel Bidirectional Encoder Representations from Transformers (BERT) model, based on Transformer encoder architecture, designed capture...
Text sentiment tendency analysis is a hot task in natural language processing. And text as the essential expression form of language, both individual word information, and overall utterance, deserves to be focused on. This paper proposes fusion model achieve high precision analysis. combines advantages CNN extract local information BiLSTM contextual association introduces attention mechanism increase focus on words with solid emotional text. The training datasets are comments that crawled...
Accurate segmentation of skin lesions is crucial for reliable clinical diagnosis and effective treatment planning. Automated techniques lesion assist dermatologists in early detection ongoing monitoring various diseases, ultimately improving patient outcomes reducing healthcare costs. To address limitations existing approaches, we introduce a novel U-shaped architecture based on our Residual Space State Block. This efficient model, termed 'SSR-UNet,' leverages bidirectional scanning to...
With the continuous development of artificial intelligence, embedding object detection algorithms into autonomous underwater detectors for marine garbage cleanup has become an emerging application area. Considering complexity environment and low resolution images taken by detectors, this paper proposes improved algorithm based on Mask R-CNN, with aim achieving high accuracy instance segmentation. First, idea dilated convolution is introduced in Feature Pyramid Network to enhance feature...
Abstract The distribution of species is not random in space. At the finest‐resolution spatial scale, that is, field sampling locations, distributional aggregation level different would be determined by various factors, for example autocorrelation or environmental filtering. However, few studies have quantitatively measured importance these factors. In this study, inspired statistical properties a Markov transition model, we propose novel additive framework to partition local multispecies...
Electric power operation, as one of the key fields in world, faces particularly prominent safety issues. Ensuring operators has become most fundamental requirement operation. However, there are some hazards construction. These mainly due to weak awareness among staff and failure standardize wearing helmets. In order effectively address this situation, technical means such video surveillance technology computer vision can be utilized monitor whether helmets provide timely feedback. Such...
At present, customers' low satisfaction and loyalty to city express service have restricted the development of express. It is particularly important analyze factors causing loyalty, which will promote industry effectively. Based on SERVQUAL model CCSI model, this paper constructs a new evaluation index system from perspective quality. Through system, first explores that affect respectively, by fuzzy analytic hierarchy process (AHP) hierarchical regression analysis, taking expected perceived...
With the development of deep learning technology, handwritten character recognition has become basic hotspot research. At same time, handwriting in minority languages is gradually worthy attention. In this paper, a robust and novel improved convolutional neural network method proposed to recognize normal digital images (Yi language). To improve accuracy fonts, there are four improvements our work. Firstly, separable convolution adopted cope with problem large parameters. Secondly, data...
To improve the accuracy of parallax maps obtained by binocular stereo matching network for better 3D reconstruction, this thesis is based on AnyNet network, with corresponding optimization to get MyNet. The Hourglass structure allows acquisition multiscale image information. original Conv replaced a residual block in avoid gradient disappearance or explosion and model degradation problems traditional convolutional neural models, thus further improving cost volume obtained. experiments were...
There is a need to improve the capability of adaptive filtering algorithm against Gaussian or multiple types non-Gaussian noises, time-varying system, and systems with low SNR. In this paper, we propose an optimized least mean absolute fourth (Optimized-LMF) algorithm, especially for unknown system signal-noise-rate (SNR). The optimal step-size LMF obtained by minimizing mean-square deviation (MSD) at any given moment in time. addition, convergence steady-state error Optimized-LMF are...
The novel coronavirus disease — COVID-19 is a historic catastrophe that has caused many devastating impacts on human life and wellness. Researchers in academia industry strive to understand the causes of this pandemic find new therapeutics combating it. Consequently, number related publications increases rapidly, it too difficult for medical researchers practitioners keep up with latest research development. Literature filtering categorization, knowledge discovery can use text mining as...
Environmental protection is still a key issue that cannot be ignored at this stage of social development. With the development artificial intelligence, various technologies increasingly tend to widely used in field environmental protection, such as searching wilderness through an unmanned aerial vehicle (UAV) and cleaning garbage by robots. Traditional object detection algorithms for scenario suffer from low accuracy high computational cost. Therefore, paper proposes algorithm applied...
Objective: Oral fluid (OF) is a new safe, non-invasive, convenient, and efficient biological sample that can be used for virus nucleic acid antibody detection. Because few studies have performed surveillance of multiple respiratory pathogens, this study sought to explore the application value OF in field. Methods: throat swabs were collected from December 2020 2021 patients with acute tract infections Beijing. Multiplex real-time PCR was performed, detection performance two samples compared....
Data non-independence, or more generally, pseudoreplication, is prevalent when biodiversity sampled using cost-effective strategies like line transects at local spatial scales with small sample sizes. Non-independence derived from sequential sampling one form of pseudoreplication. In line-transect sampling, non-independence can be well-described by a first-order Markov model. Here, we developed numerical maximum likelihood (ML) method to estimate the associated non-independent parameter for...
Objective To analyze the serotypes of enteroviruses (EVs) isolated from patients with influenza-like illness in Beijing 2017. Methods Oropharyngeal swab specimens were collected eight districts July 2017 to October 2017. EVs detected by real-time PCR. Specific primers synthesized and used amplify VP1 fragments EVs. PCR products sequenced results compared reference sequences using Basic Local Alignment Search Tool (BLAST) identify EVs. Results A total 666 91 (13.66%) positive for...
To improve the self-learning ability and information extension of existing artificial intelligence, we need a lot new intelligent theory methods to support it urgently. This paper combines intelligence technology with Extenics, puts forward operator Extension analyzes its realization path steps based on basic-element base built from multi-source knowledge graph. From Internet other open domain can extract approximately complete set data, knowledge, establish Basic-element in order through...