A5100427643- Radiomics and Machine Learning in Medical Imaging
- Advanced Neural Network Applications
- AI in cancer detection
- Domain Adaptation and Few-Shot Learning
- COVID-19 diagnosis using AI
- Medical Image Segmentation Techniques
- Medical Imaging Techniques and Applications
- Medical Imaging and Analysis
- Multimodal Machine Learning Applications
- Retinal Imaging and Analysis
- Advanced X-ray and CT Imaging
- Distributed Control Multi-Agent Systems
- Human Pose and Action Recognition
- Topic Modeling
- Cutaneous Melanoma Detection and Management
- Advanced Radiotherapy Techniques
- Digital Imaging for Blood Diseases
- Surgical Simulation and Training
- Smart Grid Security and Resilience
- Stability and Control of Uncertain Systems
- Retinal Diseases and Treatments
- Advanced Image and Video Retrieval Techniques
- Advanced Image Processing Techniques
- Artificial Intelligence in Healthcare
- Anomaly Detection Techniques and Applications
University of Hong Kong
2021-2025
Hong Kong University of Science and Technology
2021-2025
Shandong University of Traditional Chinese Medicine
2023-2025
Wenzhou Medical University
2024-2025
China Agricultural University
2011-2025
HKUST Shenzhen Research Institute
2022-2024
Guangzhou University
2024
Sichuan University
2024
Guangzhou HKUST Fok Ying Tung Research Institute
2024
Affiliated Eye Hospital of Wenzhou Medical College
2024
Liver cancer is one of the leading causes death. To assist doctors in hepatocellular carcinoma diagnosis and treatment planning, an accurate automatic liver tumor segmentation method highly demanded clinical practice. Recently, fully convolutional neural networks (FCNs), including 2-D 3-D FCNs, serve as backbone many volumetric image segmentation. However, convolutions cannot leverage spatial information along third dimension while suffer from high computational cost GPU memory consumption....
In this work, we report the set-up and results of Liver Tumor Segmentation Benchmark (LiTS), which was organized in conjunction with IEEE International Symposium on Biomedical Imaging (ISBI) 2017 Conferences Medical Image Computing Computer-Assisted Intervention (MICCAI) 2018. The image dataset is diverse contains primary secondary tumors varied sizes appearances various lesion-to-background levels (hyper-/hypo-dense), created collaboration seven hospitals research institutions. Seventy-five...
A common shortfall of supervised deep learning for medical imaging is the lack labeled data, which often expensive and time consuming to collect. This article presents a new semisupervised method image segmentation, where network optimized by weighted combination loss only inputs regularization both unlabeled data. To utilize our encourages consistent predictions network-in-training same input under different perturbations. With segmentation tasks, we introduce transformation-consistent...
Diabetic retinopathy (DR) and diabetic macular edema (DME) are the leading causes of permanent blindness in working-age population. Automatic grading DR DME helps ophthalmologists design tailored treatments to patients, thus is vital importance clinical practice. However, prior works either grade or DME, ignore correlation between its complication, i.e., DME. Moreover, location information, e.g., macula soft hard exhaust annotations, widely used as a for grading. Such annotations costly...
In this article, the event-triggered security consensus problem is studied for time-varying multiagent systems (MASs) against false data-injection attacks (FDIAs) and parameter uncertainties over a given finite horizon. process of information transmission, malicious attacker tries to inject signals destroy by compromising integrity measurements control signals. The randomly occurring stealthy FDIAs on sensors actuators are modeled Bernoulli processes. order reduce unnecessary utilization...
In this work, we report the set-up and results of Liver Tumor Segmentation Benchmark (LiTS), which was organized in conjunction with IEEE International Symposium on Biomedical Imaging (ISBI) 2017 Conferences Medical Image Computing Computer-Assisted Intervention (MICCAI) 2018. The image dataset is diverse contains primary secondary tumors varied sizes appearances various lesion-to-background levels (hyper-/hypo-dense), created collaboration seven hospitals research institutions. Seventy-five...
Generalizing the medical image segmentation algorithms to unseen domains is an important research topic for computer-aided diagnosis and surgery. Most existing methods require a fully labeled dataset in each source domain. Although some researchers developed semi-supervised domain generalized method, it still requires labels. This paper presents novel confidence-aware cross pseudo supervision algorithm segmentation. The main goal enhance label quality unlabeled images from unknown...
Liver cancer is one of the leading causes death. To assist doctors in hepatocellular carcinoma diagnosis and treatment planning, an accurate automatic liver tumor segmentation method highly demanded clinical practice. Recently, fully convolutional neural networks (FCNs), including 2D 3D FCNs, serve as back-bone many volumetric image segmentation. However, convolutions can not leverage spatial information along third dimension while suffer from high computational cost GPU memory consumption....
The automatic diagnosis of various retinal diseases from fundus images is important to support clinical decision-making. However, developing such solutions challenging due the requirement a large amount human-annotated data. Recently, unsupervised/self-supervised feature learning techniques receive lot attention, as they do not need massive annotations. Most current self-supervised methods are analyzed with single imaging modality and there no method currently utilize multi-modal for better...
Computed tomography (CT) has been widely used for medical diagnosis, assessment, and therapy planning guidance. In reality, CT images may be affected adversely in the presence of metallic objects, which could lead to severe metal artifacts influence clinical diagnosis or dose calculation radiation therapy. this article, we propose a generalizable framework artifact reduction (MAR) by simultaneously leveraging advantages image domain sinogram domain-based MAR techniques. We formulate our as...
Automatic skin lesion segmentation on dermoscopic images is an essential component in computer-aided diagnosis of melanoma. Recently, many fully supervised deep learning based methods have been proposed for automatic segmentation. However, these approaches require massive pixel-wise annotation from experienced dermatologists, which very costly and time-consuming. In this paper, we present a novel semi-supervised method by leveraging both labeled unlabeled data. The network optimized the...
The problem of finite-horizon H <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">∞</sub> state estimator design for periodic neural networks over multiple fading channels is studied in this paper. To characterize the measurement signals transmitted through different experiencing channel fading, a model considered. For investigating situation correlated channels, set random variables introduced. Specifically, coefficients are described by white...
The automatic diagnosis of various conventional ophthalmic diseases from fundus images is important in clinical practice. However, developing such solutions challenging due to the requirement a large amount training data and expensive annotations for medical images. This paper presents novel self-supervised learning framework retinal disease reduce annotation efforts by visual features unlabeled To achieve this, we present rotation-oriented collaborative method that explores rotation-related...
Federated semi-supervised learning (FSSL) aims to derive a global model by training fully-labeled and fully-unlabeled clients or partially labeled clients. The existing approaches work well when local have in-dependent identically distributed (IID) data but fail generalize more practical FSSL setting, i.e., Non-IID setting. In this paper, we present Random Sampling Consensus learning, namely RSCFed, con-sidering the uneven reliability among models from clients, Our key motivation is that...
Weakly-Supervised Semantic Segmentation (WSSS) segments objects without heavy burden of dense annotation. While as a price, generated pseudo-masks exist obvious noisy pixels, which result in sub-optimal segmentation models trained over these pseudo-masks. But rare studies notice or work on this problem, even pixels are inevitable after their improvements pseudo-mask. So we try to improve WSSS the aspect noise mitigation. And observe that many high confidences, especially when response range...
Contrastive Language-Image Pre-training (CLIP) is a powerful multimodal large vision model that has demonstrated significant benefits for downstream tasks, including many zero-shot learning and text-guided tasks. However, we notice some severe problems regarding the model's explainability, which undermines its credibility impedes related Specifically, find CLIP prefers background regions than foregrounds according to predicted similarity map, contradicts human understanding. Besides, there...
Dissolved organic matter (DOM) sustains a substantial part of the transported seaward, where photochemical reactions significantly affect its transformation and fate. The irradiation experiments can provide valuable information on reactivity (photolabile, photoresistant, photoproduct) molecules. However, inconsistency fate irradiated molecules among different curtailed our understanding roles have played, which cannot be properly addressed by traditional approaches. Here, we conducted for...