A5102893703- Human Pose and Action Recognition
- Diabetes Treatment and Management
- Multimodal Machine Learning Applications
- Gait Recognition and Analysis
- Acupuncture Treatment Research Studies
- Statistical Methods and Bayesian Inference
- Context-Aware Activity Recognition Systems
- Advanced Image and Video Retrieval Techniques
- Pancreatic function and diabetes
- Industrial Vision Systems and Defect Detection
- Domain Adaptation and Few-Shot Learning
- Image Processing and 3D Reconstruction
- Statistical Methods in Clinical Trials
- Vehicle License Plate Recognition
- Advanced MIMO Systems Optimization
- Multiple Myeloma Research and Treatments
- Power Line Communications and Noise
- Statistical Methods and Inference
- Medicinal plant effects and applications
- Apelin-related biomedical research
- Complementary and Alternative Medicine Studies
- Advanced Measurement and Detection Methods
- Lymphoma Diagnosis and Treatment
- Drug Solubulity and Delivery Systems
- Meta-analysis and systematic reviews
Institute of Automation
2020-2024
Chinese Academy of Sciences
2020-2024
First People's Hospital of Chongqing
2023
Brookhaven National Laboratory
2023
University of Chinese Academy of Sciences
2020-2022
Peng Cheng Laboratory
2020-2021
Beijing Academy of Artificial Intelligence
2021
University of Maryland, Baltimore County
2010-2020
National Tsing Hua University
2019
Ordnance Engineering College
2016
Unpaired image-to-image translation has broad applications in art, design, and scientific simulations. One early breakthrough was CycleGAN that emphasizes one-to-one mappings between two unpaired image domains via generative-adversarial networks (GAN) coupled with the cycle-consistency constraint, while more recent works promote one-to-many mapping to boost diversity of translated images. Motivated by simulation needs, this work revisits classic framework boosts its performance outperform...
Federated human activity recognition (FHAR) has attracted much attention due to its great potential in privacy protection. Existing FHAR methods can collaboratively learn a global model based on unimodal or multimodal data distributed different local clients. However, it is still questionable whether existing work well more common scenario where are from modalities, e.g., some clients may provide motion signals while others only visual data. In this paper, we study new problem of cross-modal...
Deep neural networks(DNN) is an important method for machine learning, which has been widely used in many fields. Compared with the shallow networks(NN), DNN better feature expression and ability to fit complex mapping. In this paper, we first introduce background of development DNN, then several typical model, including deep belief networks(DBN), stacked autoencoder(SAE) convolution networks(DCNN), finally research its applications from three aspects prospects direction DNN.
Recognizing activities from egocentric multimodal data collected by wearable cameras and sensors, is gaining interest, as methods always benefit the complementarity of different modalities. However, since high-dimensional videos contain rich high-level semantic information while low-dimensional sensor signals describe simple motion patterns wearer, large modality gap between raises a challenge for fusing raw data. Moreover, lack large-scale datasets due to cost collection annotation...
Video domain adaptation aims to transfer knowledge from labeled source videos unlabeled target videos. Existing video methods require full access the reduce gap between and videos, which are impractical in real scenarios where not available with concerns transmission efficiency or privacy issues. To address this problem, paper, we propose solve a source-free task for only pre-trained model learning multimodal classification model. cannot be directly applied task, since always suffer...
Abstract Traditional Chinese medicine (TCM), used in China and other Asian counties for thousands of years, is increasingly utilized Western countries. However, due to inherent differences how this ancient modality are practiced, employing the so‐called medicine‐based gold standard research methods evaluate TCM challenging. This paper a discussion obstacles design statistical analysis clinical trials TCM. It based on our experience designing conducting randomized controlled trial acupuncture...
Introduction: The efficacy of dapagliflozin remains controversial for patients with type 2 diabetes and non-alcoholic fatty liver disease. We conduct this meta-analysis to explore the influence versus placebo on treatment complicated disease.
 Methods: have searched PubMed, EMbase, Web science, EBSCO, Cochrane library databases through November 2021 randomized controlled trials (RCTs) assessing This is performed using random-effect model.
 Results: Four RCTs are included in...
Empirical likelihood-based confidence intervals for the population mean have many interesting properties [Owen, A.B. (1988), 'Empirical Likelihood Ratio Confidence Intervals a Single Functional', Biometrika, 75, 237–249]. Calibrated by χ2 limiting distribution, however, their coverage probabilities are often lower than nominal when sample size is small and/or dimension of data high. The application adjusted empirical likelihood (AEL) one ways to achieve more accurate probability. In this...
Though existing cross-domain action recognition methods successfully improve the performance on videos of one view ( <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">e.g.</i> , egocentric videos) by transferring knowledge from another exocentric videos), they have limitations in generality because source and target domains need to be fixed aforehand. In this paper, we propose solve a more practical task multi-domain egocentric-exocentric...
In this paper, we consider the task of action anticipation on egocentric videos. Previous methods ignore explicit modeling global context relation among past and future actions, which is not an easy due to vacancy unobserved To solve problem, propose a Multimodal Global Relation Knowledge Distillation (MGRKD) framework distill knowledge learned from full videos improve partially observed The proposed MGRKD has teacher-student learning strategy, where either teacher or student model three...
Knee osteoarthritis (OA) manifests with pain, joint stiffness, and limited function. In traditional Chinese medicine, knee OA is differentiated into three patterns: yang deficiency cold coagulation, kidney deficiency, blood stasis. The objective of this study was to determine whether coagulation patients respond better thermal laser acupuncture treatment than do non-yang deficient patients. Fifty-two were allocated group A (yang deficient, n = 26) or B (non-yang 26). All received a 20-min at...
Meaningful comparison of the dissolution profiles between reference and test formulations a drug is critical for assessing similarity two formulations, quality control purposes. Such profile required by regulatory authorities, criteria used this include widely difference factor f1 f2 , recommended Food Drug Administration . In spite their extensive use in practice, factors have been heavily criticized on various grounds; criticisms ignoring sampling variability correlations across time...