A5091309588- Advanced Text Analysis Techniques
- Topic Modeling
- Web Data Mining and Analysis
- Advanced Neural Network Applications
- Domain Adaptation and Few-Shot Learning
- Advanced Computational Techniques and Applications
- Face and Expression Recognition
- Parkinson's Disease Mechanisms and Treatments
- Advanced Image and Video Retrieval Techniques
- Machine Learning and ELM
- Information Retrieval and Search Behavior
- Smart Agriculture and AI
- Text and Document Classification Technologies
- RFID technology advancements
- Neurological disorders and treatments
- Sentiment Analysis and Opinion Mining
- Fault Detection and Control Systems
- Image Retrieval and Classification Techniques
- Geoscience and Mining Technology
- Educational Technology and Pedagogy
- Energetic Materials and Combustion
- Manufacturing Process and Optimization
- Natural product bioactivities and synthesis
- Emotion and Mood Recognition
- Visual Attention and Saliency Detection
Henan University of Science and Technology
2016-2025
Command Hospital
2020
General Hospital of Shenyang Military Region
2009-2020
Modern Electron (United States)
2014
Luoyang Institute of Science and Technology
2010-2013
Institute of Automation
2010
Shanghai Jiao Tong University
2005-2009
Northwestern Polytechnical University
2008
Harbin University
2007
Harbin Engineering University
2007
Parkinson's disease is a neurological disorder, and early diagnosis crucial for the treatment quality of life patients. Gait movement disorder significant manifestation PD, automated gait assessment key to achieving detection PD With development deep learning, in order improve accuracy enhance robustness motion recognition models, this study introduces an innovative learning approach, namely Multi-area Attention Spatiotemporal Directed Graph Convolutional Network (Ma-ST-DGN). The model...
In the rapidly evolving field of computer vision and machine learning, 3D skeleton estimation is critical for applications such as motion analysis human–computer interaction. While stereo cameras are commonly used to acquire skeletal data, monocular RGB systems attract attention due benefits including cost-effectiveness simple deployment. However, persistent challenges remain in accurately inferring depth from 2D images reconstructing structures using approaches. The current methods overly...
The diagnosis of Parkinson’s disease relies heavily on the subjective assessment physicians, which depends their individual experience and training, potentially leading to inconsistent diagnostic results. Therefore, developing an objective efficient method is essential improve accuracy timeliness diagnosis. In this study, we utilized PhysioNet dataset, a time-series dataset comprising data from 93 patients 73 healthy individuals. contains vertical ground reaction forces recorded 16 sensors...
Abstract Maize leaf disease seriously affects maize yield, a identification model with an improved lightweight network EfficientNet was proposed in this study. First, the replaces SENet module MBConv CBAM module, so that not only focuses on correlation between channels but also adaptively learns attentional weight of each spatial location. Furthermore, multi‐scale feature fusion layer based residual connection is introduced to extract more comprehensive and richer features at different...
A single network model can't extract more complex and rich effective features. Meanwhile, the structure is usually huge, there are many parameters consume space resources, etc. Therefore, combination of multiple models to complementary features has attracted extensive attention. In order solve problems existing in prior art that high spatial depth features, redundant parameters, weak generalization ability, this paper adopts two Xception module inverted residual build neural network. Based...
In the last 70 years, automatic text summarization work has become more and important because amount of data on Internet is increasing so fast, can extract useful information knowledge what user's need that could be easily handled by humans used for many purposes. Especially in people's daily life, news type most people are exposed to. this study, a new summarzation model which based fuzzy logic rules, multi-feature Genetic algorithm (GA) introduced. Firstly, feature word features, we score...
To compare the differences between asleep and awake robot-assisted deep brain stimulation (DBS) surgery for Parkinson's Disease (PD), we conducted this retrospective cohort study included 153 PD patients undergoing bilateral DBS from June 2017 to August 2019, of which 58 cases were performed under general anesthesia (GA) 95 local (LA). Procedure duration, parameters, electrode implantation accuracy, intracranial air, intraoperative electrophysiological signal length, complications, Unified...
The emergence of CAPTCHA ( Completely Automated Public Turing Test to Tell Computers and Humans Apart )is better protect the network security ,and research on recognition technology is conducive expand design ideas improve loopholes original design.In order rate CAPTCHA, we proposed a method with Adhesion Character, which effectively improves segmentation quality adhesive character in complex background.First based preprocessing image utilize connected area noise reduction further denoise...
Text sentiment analysis plays an important role in social network information mining.It is also the theoretical foundation and basis of personalized recommendation, circle interest classification public opinion analysis.In view existing algorithms for feature extraction weight calculation, we find that they fail to fully take into account influence words.Therefore, this paper proposed a fine-grained short text method based on machine learning.To improve calculation selection weighting more...
Fine-grained image classification is a sub-category problem with common superior category. Aiming at the characteristics of large intra-class differences and small inter-class in fine-grained images, this paper proposes method based on multi-scale feature fusion. The constructs three-branch network model. attention module local extraction are used to obtain target object parts strong distinguishing detail features. depth metric learning shorten distance from same data by using...
Nowadays, the field of video-based action recognition is rapidly developing. Although Vision Transformers (ViT) have made great progress in static image processing, they are not yet fully optimized for dynamic video applications. Convolutional Neural Networks (CNN) and related models perform exceptionally well recognition. However, there still some issues that cannot be ignored, such as high computational costs large memory consumption. In face these issues, current research focuses on...
Printed circuit board (PCB) defect detection is an important and indispensable part of industrial production. PCB defects, due to the small target similarity between classes, in actual production process are prone omission false problems. Traditional machine-learning-based methods limited by needs do not show good results. Aiming at problems related detection, we propose a algorithm based on DSASPP-YOLOv5 conduct experiments PKU-Market-PCB dataset. improved single-stage model, first used...
The feature of large intra-class variance in fine-grained image classification is a challenge to the task. How effectively learn discriminant objects graph and find out small regions key classification. This paper proposes weak-supervised algorithm based on multi-granularity fusion. ECA module fused with classic network ResNet-50 optimize residual block obtain new basic enhance channel attention. Secondly, local chaos introduced into form through random regrouping so that can different...