A5100319145- Balance, Gait, and Falls Prevention
- Cerebral Palsy and Movement Disorders
- Advanced Image and Video Retrieval Techniques
- Multimodal Machine Learning Applications
- Advanced Neural Network Applications
- Video Analysis and Summarization
- Muscle activation and electromyography studies
- Video Surveillance and Tracking Methods
- Topic Modeling
- Human Pose and Action Recognition
- Domain Adaptation and Few-Shot Learning
- Advanced Text Analysis Techniques
- Image Retrieval and Classification Techniques
- Stroke Rehabilitation and Recovery
- Robotics and Sensor-Based Localization
- Image Enhancement Techniques
- Face recognition and analysis
- Higher Education and Teaching Methods
- Context-Aware Activity Recognition Systems
- Natural Language Processing Techniques
- EEG and Brain-Computer Interfaces
- Underwater Acoustics Research
- Anomaly Detection Techniques and Applications
- Face and Expression Recognition
- Sentiment Analysis and Opinion Mining
Czech Academy of Sciences, Institute of Physics
2025
Chinese Academy of Sciences
2025
Nanning Normal University
2025
Tencent (China)
2020-2024
Minzu University of China
2024
Beijing Institute of Technology
2018-2023
Shenzhen Institutes of Advanced Technology
2023
Chengdu University of Technology
2022-2023
Xiamen University
2022
University of Illinois Chicago
2018-2022
Rotated object detection aims to identify and locate objects in images with arbitrary orientation. In this scenario, the oriented directions of vary considerably across different images, while multiple orientations exist within an image. This intrinsic characteristic makes it challenging for standard backbone networks extract high-quality features these arbitrarily orientated objects. paper, we present Adaptive Convolution (ARC) module handle afore-mentioned challenges. our ARC module,...
Cross domain object detection is a realistic and challenging task in the wild. It suffers from performance degradation due to large shift of data distributions lack instance-level annotations target domain. Existing approaches mainly focus on either these two difficulties, even though they are closely coupled cross detection. To solve this problem, we propose novel Target-perceived Dual-branch Distillation (TDD) framework. By integrating branches both source domains unified teacher-student...
Attribute recognition is a crucial but challenging task due to viewpoint changes, illumination variations and appearance diversities, etc. Most of previous work only consider the attribute-level feature embedding, which might perform poorly in complicated heterogeneous conditions. To address this problem, we propose hierarchical embedding (HFE) framework, learns fine-grained by combining attribute ID information. In HFE, maintain inter-class intra-class simultaneously. Not samples with same...
With the change of education concept and development English teaching, cultivating students' autonomous learning ability has gradually become one important goals teaching. Taking oral as an example, this paper discusses how to cultivate in Eng-lish course From perspective theory practice, comprehensively analyzes concept, characteristics importance au-tonomous learning, puts forward some strategies for learning. Based on relevant research empirical analysis, study designed em-pirical explore...
<title>Abstract</title> Antiferromagnetism has become a promising candidate for the next generation electronic devices due to its thermal stability, low energy consumption, and fast switching speed. However, canceling of net magnetic moment in antiferromagnetic order presents great challenge on quantitative characterization modulation, hindering investigation application. In this work, utilizing optical second harmonic (SHG) wide temperature range, integrated differential phase contrast...
Electroencephalography (EEG) can reveal the abnormalities of dopaminergic subcortico-cortical circuits in patients with Parkinson's disease (PD). However, conventional time-frequency analysis EEG signals cannot fully non-linear processes neural activities and interactions. A novel Holo-Hilbert Spectral Analysis (HHSA) was applied to features resting state 99 PD 59 healthy controls (HCs). demonstrated a reduction β bands frontal central regions, γ central, parietal, temporal regions. Compared...
Large scale surveillance video analysis is one of the most important components in future artificial intelligent city. It a very challenging but practical system, consists multiple functionalities such as object detection, tracking, identification and behavior analysis. In this paper, we try to address three tasks hosted NVIDIA AI City Challenge contest. First, system that transforming image coordinate world has been proposed, which useful estimate vehicle speed on road. Second, anomalies...
The purpose was to examine and compare the longer-term generalization between 2 different practice dosages for a single-session treadmill slip-perturbation training when reexposed an overground slip 6 months later. A total of 45 older adults were conveniently assigned either 24 or 40 slip-like perturbation trials third control group. Overground slips given immediately after initial training, at in order immediate effects. performance (center mass stability vertical limb support) fall...
This study examined the effects of perturbation training on contextual interference and generalization encountering a novel opposing perturbation. One hundred sixty-nine community-dwelling healthy older adults (69.6 ± 6.4 years) were randomly assigned to one three groups: slip-perturbation (St, n = 67) group received 24 slips, trip-perturbation (Tt, trips, control (Ctrl: 31) only non-perturbed walking trials ( ClinicalTrials.gov NCT03199729; https://clinicaltrials.gov/ct2/show/NCT03199729 )....
The task of cross-modal image retrieval has recently attracted considerable research attention. In real-world scenarios, keyword-based queries issued by users are usually short and have broad semantics. Therefore, semantic diversity is as important accuracy in such user-oriented services, which improves user experience. However, most typical methods based on single point query embedding inevitably result low diversity, while existing diverse approaches frequently lead to due a lack...
Slip-induced falls are a growing health concern for older adults, and near-fall events associated with an increased risk of falling. To detect adults at high slip-related falls, this study aimed to develop models event detection based on accelerometry data collected by body-fixed sensors. Thirty-four healthy who experienced 24 laboratory-induced slips were included. The slip outcomes first identified as loss balance (LOB) no LOB (NLOB), then the kinematic measures compared between these two...
Slip outcomes are categorized as either a backward loss of balance (LOB) or no (no-LOB) in which an individual does not take step to regain their stability. LOB includes falls and nonfalls, while no-LOB skate overs walkovers. Researchers uncertain about factors determine slip at critical instants they do so. The purpose the study was investigate affecting proactive early reactive phases by analyzing 136 trials from 68 participants (age: 72.2 [5.3] y, female: 22). Segment angles average joint...
Text coherence plays a key role in document quality assessment. Most existing text methods only focus on similarity of adjacent sentences. However, local exists sentences with broader contexts and diverse rhetoric relations, rather than just similarity. Besides, the highlevel is also an important aspect quality. To this end, we propose hierarchical model for In our model, implement attention mechanism to capture location semantics, bilinear tensor layer measure max-coherence pooling...
Oriented object detection, an emerging task in recent years, aims to identify and locate objects across varied orientations. This requires the detector accurately capture orientation information, which varies significantly within images. Despite existing substantial efforts, simultaneously ensuring model effectiveness parameter efficiency remains challenging this scenario. In paper, we propose a lightweight yet effective Group-wise Rotating Attention (GRA) module replace convolution...
Rotated object detection aims to identify and locate objects in images with arbitrary orientation. In this scenario, the oriented directions of vary considerably across different images, while multiple orientations exist within an image. This intrinsic characteristic makes it challenging for standard backbone networks extract high-quality features these arbitrarily orientated objects. paper, we present Adaptive Convolution (ARC) module handle aforementioned challenges. our ARC module,...