A5100689167- Advanced Vision and Imaging
- Text and Document Classification Technologies
- Advanced Computational Techniques and Applications
- Anomaly Detection Techniques and Applications
- Data Management and Algorithms
- Advanced Neural Network Applications
- Neuroscience and Neuropharmacology Research
- Optical measurement and interference techniques
- Advanced Text Analysis Techniques
- Neural Networks and Applications
- Ion channel regulation and function
- AI in cancer detection
- Bayesian Modeling and Causal Inference
- Advanced Clustering Algorithms Research
- Web Data Mining and Analysis
- Face and Expression Recognition
- Advanced Graph Neural Networks
- Image Processing Techniques and Applications
- Breast Lesions and Carcinomas
- Robotics and Sensor-Based Localization
- Topic Modeling
- Rough Sets and Fuzzy Logic
- Blind Source Separation Techniques
- Global Health and Surgery
- Image and Video Stabilization
Yanshan University
2025
Peking University
2015-2024
Beijing Jiaotong University
2014-2024
Shanghai First People's Hospital
2024
Dalian University of Technology
2020-2024
Dalian University
2024
Capital Medical University
2023
Beijing Chao-Yang Hospital
2023
Guangzhou University of Chinese Medicine
2022
Jilin Agricultural University
2010-2021
The Lancet Commission on Global Surgery established the Three Delays framework, categorising delays in accessing timely surgical care into seeking (First Delay), reaching (Second and receiving (Third Delay). Globally, knowledge gaps regarding for fracture care, lack of large prospective studies informed rationale our international observational study. We investigated hospital admission as a surrogate explored factors associated with delayed admission.
3D object detection is an important scene understanding task in autonomous driving and virtual reality. Approaches based on LiDAR technology have high performance, but expensive. Considering more general scenes, where there no data the datasets, we propose a approach from stereo vision which does not rely either as input or supervision training, solely takes RGB images with corresponding annotated bounding boxes training data. As depth estimation of key factor affecting performance...
In response to the constraints of current methods in swiftly and accurately identifying complex flow patterns under actual conditions, this paper proposes an innovative identification method for oil-water two-phase patterns, based on a Spatial Scale Internal attention feature Fusion Network (SSIFNet). Specially, spatial scale module is designed equip model with scale-aware ability capture characteristics at varying scales patterns. Moreover, internal strategy developed realize local context...
Multi-document summarization is of great value to many real world applications since it can help people get the main ideas within a short time.In this paper, we tackle problem extracting summary sentences from multi-document sets by applying sparse coding techniques and present novel framework challenging problem. Based on data reconstruction sentence denoising assumption, two-level representation model depict process summarization. Three requisite properties proposed form an ideal...
Remote sensing cross-modal text-image retrieval (RSCTIR) has recently attracted extensive attention due to its advantages of fast extraction remote image information and flexible human–computer interaction. Traditional RSCTIR methods mainly focus on improving the performance uni-modal feature separately, most rely pre-trained object detectors obtain better local representation, which not only lack multi-modal interaction information, but also cause training gap between detector task. In this...
Deep learning-based computer-aided diagnosis (CAD) is an important method in aiding for radiologists. We investigated the accuracy of a deep CAD classifying breast lesions with different histological types.A total 448 were detected on ultrasound (US) and classified by experienced radiologist, resident respectively. The pathological results chosen as golden standard. diagnostic performances three raters types analyzed.For overall performance, presented significantly higher specificity...
Pentameric ligand-gated ion channels (LGICs) play conserved, critical roles in both excitatory and inhibitory synaptic transmission can be activated by diverse neurochemical ligands. We have performed a characterization of orphan from the nematode C. elegans, identifying five new monoamine-gated LGICs with functional properties expression postsynaptic to aminergic neurons. These include polymodal anion dopamine tyramine, which may mediate molecules vivo. Intriguingly, we also find that novel...
This paper mainly completes the binary classification of RCV1 text data set by logistic regression. Based on established regression model, performance and characteristics three numerical optimization algorithms–random gradient descent, Mini-Batch random L-BFGS are studied.
Independent component analysis (ICA), instead of the traditional discrete cosine transform (DCT), is often used to project log Mel spectrum in robust speech feature extraction. The paper proposed using symmetric orthogonalization ICA for projecting into a new space as substitute extracting features solve problem cumulative error and unequal weights that deflation brings, so improve robustness recognition systems, increase efficiency estimation at same time. Furthermore, studied...
Outlier detection, as a type of one-class classification problem, is one important research topics in data mining and machine learning. Its task to identify sample points markedly deviating from the normal data. A reliable outlier detector needs build model which encloses tightly. In this paper, an improved SVM (OC-SVM) classifier proposed for detection problems. We name method OC-SVM with minimum within-class scatter (OC-WCSSVM), exploits inner-class structure training set via minimizing...
Abstract The sodium channel β1 subunit affects gating and surface density, but little is known about the factors that regulate expression or its participation in fine control of cellular excitability. In this study we examined whether graded contributes to gradient current inactivation, which tightly controlled directly related a social behavior, electric organ discharge (EOD), weakly fish Sternopygus macrurus . We found mRNA protein levels both correlate with EOD frequency. identified novel...
The performance of traditional mel-frequency cepstral coefficients (MFCC) speech feature extraction method decreases drastically in the complex noisy environment. To improve and robustness recognition system, which is based on spectral envelope estimation method, minimum distortionless response spectrum MVDR-MFCC (Minimum Variance Distortionless Response-MFCC) was proposed. However, computational complexity very high. In this paper, we proposed MHCC (Hilbert-MFCC) for speech, introduced...
提出了一种半监督K均值多关系数据聚类算法.该算法在K均值聚类算法的基础上扩展了其初始类簇的选择方法和对象相似性度量方法,以用于多关系数据的半监督学习.为了获取高性能,该算法在聚类过程中充分利用了标记数据、对象属性及各种关系信息.多关系数据库Movie上的实验结果验证了该算法的有效性.;A semi-supervised K-means clustering algorithm for multi-type relational data is proposed, which extends traditional by new methods of selecting initial clusters and similarity measures, so that it can semi-supervise cluster data. In order to achieve high performance, in the algorithm, besides attribute information, both labeled relationship...
As a common breast cancer-related complaint, pathological nipple discharge (PND) detected by ductoscopy is often missed diagnosed. Deep learning techniques have enabled great advances in clinical imaging but are rarely applied cancer with PND. This study aimed to design and validate an Intelligent Ductoscopy for Breast Cancer Diagnostic System (IDBCS) diagnosis analyzing real-time data acquired ductoscopy.The present multicenter, case-control trial was carried out 6 hospitals China. Images...
Adversarial examples are carefully perturbed input that aim to mislead the deep neural network models into producing unexpected outputs. In this paper, we employ a K-means clustering algorithm as pre-processing method defend against adversarial examples. Specifically, reconstruct according their cluster assignments in pixel level reduce impact of injected perturbation. Our approach does not rely on any architectures and can also work with existing defenses provide better protection for...
The influence of online public opinion on agricultural product safety the society is increasing. In order to correctly guide direction products, help sector turn from passive active opinion, timely prevent spread negative and reduce impact hot events, it especially important improve ability monitoring products’ network opinion. This research based big data technology develop an system that can collect, process, analyze in real time, discover track topics, calculate visualize polarity...
Net traffic classification is the basis for providing various network services such as security, monitoring and Quality of Service (QoS) etc. Therefore, this field has always been a hot spot in academic industrial research. Through proper data processing, researcher said that it possible to classify flows through machine learning. necessary explore suitable processing methods model structures. To our best knowledge, based on sequential feature learning are rarely discussed, so paper proposes...