A5046390476- Molecular Biology Techniques and Applications
- Advanced MRI Techniques and Applications
- RNA Research and Splicing
- Metabolomics and Mass Spectrometry Studies
- Medical Image Segmentation Techniques
- Robotic Path Planning Algorithms
- Robotics and Automated Systems
- Advanced Malware Detection Techniques
- Advanced Neuroimaging Techniques and Applications
- Speech and Audio Processing
- Dengue and Mosquito Control Research
- Topic Modeling
- Cell Image Analysis Techniques
- Medical Imaging Techniques and Applications
- Congenital heart defects research
- Cardiomyopathy and Myosin Studies
- Autonomous Vehicle Technology and Safety
- Augmented Reality Applications
- Surgical Simulation and Training
- Visual Attention and Saliency Detection
- Advanced Image Processing Techniques
- Image Enhancement Techniques
- Stochastic Gradient Optimization Techniques
- Radiomics and Machine Learning in Medical Imaging
- Protein Kinase Regulation and GTPase Signaling
University of Hong Kong
2024-2025
Zhejiang University-University of Edinburgh Institute
2023-2024
Chinese University of Hong Kong
2024
South China University of Technology
2024
Beijing University of Posts and Telecommunications
2023-2024
Hospital for Sick Children
2004-2011
University of Toronto
2004-2009
SickKids Foundation
2004-2007
Great Ormond Street Hospital
2004
University College London
2004
High-frequency ultrasound biomicroscopy (UBM) has recently emerged as a high-resolution means of phenotyping genetically altered mice and great potential to evaluate the cardiac morphology hemodynamics mouse mutants. However, there is no standard procedure in vivo transthoracic imaging using UBM comprehensively phenotype adult mice. In this paper, characteristic thoracic anatomy elucidated magnetic resonance (MR) on fixed Besides left parasternal apical windows commonly used for imaging,...
Deep Reinforcement Learning (DRL) suffers from uncertainties and inaccuracies in the observation signal realworld applications. Adversarial attack is an effective method for evaluating robustness of DRL agents. However, existing methods targeting individual sampled actions have limited impacts on overall policy distribution, particularly continuous action spaces. To address these limitations, we propose Distribution-Aware Projected Gradient Descent (DAPGD). DAPGD uses distribution similarity...
Abstract The use of mice to study models human disease has resulted in a surge interest developing mouse MRI. ability take 3D, high‐resolution images live allows significant insight into anatomy and function. However, with imaging times on the order hours, high throughput specimens been problematic. To facilitate throughput, concurrent multiple developed; however, this poses further complexities regarding ease rapidity loading several animals. In study, custom‐built equipment was developed...
Abstract Impaired cognitive, memory, or motor performance is a distinguishing characteristic of neurological diseases. Although these symptoms are frequently the most evident in human patients, additional markers disease critical for proper diagnosis and staging. Noninvasive neuroimaging methods have become essential this capacity provide means evaluating tracking progression. These imaging also becoming available to scientists research laboratory assessment animal models disease. Imaging...
Abstract High‐resolution magnetic resonance imaging is emerging as a powerful tool for phenotyping mice in biologic studies of genetic expression, development, and disease progression. In several applications, notably random mutagenesis trials, large cohorts must be examined abnormalities that may occur any part the body. aim establishing protocol multiple simultaneously standardized high‐throughput fashion, this study investigates variations three‐dimensional fast spin‐echo sequence...
Abstract With a fixed time to acquire magnetic resonance (MR) image, can be spent better spatial resolution with decrease in signal‐to‐noise ratio (SNR) or decreased increase SNR. This resolution/SNR tradeoff at has been investigated by visual rater study using images of ex vivo mouse brains. Simulated between SNR and were produced from high‐quality, 3D isotropic brain emulate shorter constant acquisition times. The spanned range SNRs (63–6) resolutions (32–81 µm). Fourteen readers...
Abstract Image registration serves many applications in medical imaging, including longitudinal studies, treatment verification, and more recently, morphometry. Registration processing is regularly applied magnetic resonance (MR) images, where imaging highly adaptable capturing soft tissue contrast. To obtain the greatest accuracy MR inherent tradeoff between SNR resolution at a given scan time should be optimized for computational accuracy, rather than human viewing. We investigated this...
Recently, there has been a surge in the development of advanced intelligent generative content (AIGC), especially large language models (LLMs). However, for many downstream tasks, it is necessary to fine-tune LLMs using private data. While federated learning offers promising privacy-preserving solution LLM fine-tuning, substantial size an LLM, combined with high computational and communication demands, makes hard apply tasks. More importantly, edge servers often possess varying computing...
Monitoring sleeping postures is a critical task in the healthcare field, as improper positions can result issues such pressure ulcers and sleep apnea. Furthermore, specific must be maintained by certain patients to facilitate better postoperative recovery. Currently, this primarily relies on manual monitoring questioning, which labor-intensive susceptible subjective errors. In study, we propose an algorithm for pose estimation based Skinned Multi-Person Linear (SMPL) model, accurately...
Recently neural radiance fields (NeRF) have been widely exploited as 3D representations for dense simultaneous localization and mapping (SLAM). Despite their notable successes in surface modeling novel view synthesis, existing NeRF-based methods are hindered by computationally intensive time-consuming volume rendering pipeline. This paper presents an efficient RGB-D SLAM system, i.e., CG-SLAM, based on a uncertainty-aware Gaussian field with high consistency geometric stability. Through...
Voice authentication has been widely used on smartphones. However, it remains vulnerable to spoofing attacks, where the attacker replays recorded voice samples from authentic humans using loudspeakers bypass system. In this paper, we present MagLive, a robust liveness detection scheme designed for smartphones mitigate such attacks. MagLive leverages differences in magnetic field patterns generated by different speakers (i.e., or loudspeakers) when speaking detection. It uses built-in...
Cooperative perception enhances sensing in multi-robot and vehicular networks by aggregating information from multiple agents, improving accuracy range. However, mobility non-rigid sensor mounts introduce extrinsic calibration errors, necessitating online calibration, which is complicated limited overlap regions. Maintaining fresh crucial for timely accurate sensing. To address errors ensure both transmission timeliness, we propose a Robust Task-Oriented Communication framework (R-TOCOM)...