A5100360466- Radiomics and Machine Learning in Medical Imaging
- Face and Expression Recognition
- AI in cancer detection
- Medical Image Segmentation Techniques
- Medical Imaging Techniques and Applications
- Image Retrieval and Classification Techniques
- Topic Modeling
- Advanced Image and Video Retrieval Techniques
- Pancreatic and Hepatic Oncology Research
- Remote-Sensing Image Classification
- Domain Adaptation and Few-Shot Learning
- Natural Language Processing Techniques
- Video Surveillance and Tracking Methods
- Machine Learning and ELM
- Multimodal Machine Learning Applications
- Advanced X-ray and CT Imaging
- Advanced Image Processing Techniques
- Image Processing Techniques and Applications
- Neural Networks and Applications
- Image and Signal Denoising Methods
- Advanced Image Fusion Techniques
- Advanced Computing and Algorithms
- Advanced Vision and Imaging
- Lung Cancer Treatments and Mutations
- Biometric Identification and Security
University of Macau
2019-2025
Tencent (China)
2024
City University of Macau
2023-2024
Anhui University
2024
Shanghai University
2015-2024
Chengdu University
2024
Harbin University of Science and Technology
2023
Microsoft (Finland)
2022
Shanghai Institute of Technology
2022
University of Science and Technology
2022
Solid-state electrolytes are currently receiving increasing interest due to their high mechanical strength and chemical stability for safe battery construction. However, poor ion conduction unclear mechanism need further improvement exploration. This study focuses on a hybrid solid-state electrolyte containing MOF-based scaffolds, using metal salts as the conductor. In this paper, we employ an substitution strategy manipulate scaffold structure at lattice level by replacing hydrogen with...
The clustering-based unsupervised relation discovery method has gradually become one of the important methods open extraction (OpenRE). However, high-dimensional vectors can encode complex linguistic information which leads to problem that derived clusters cannot explicitly align with relational semantic classes. In this work, we propose a relation-oriented clustering model and use it identify novel relations in unlabeled data. Specifically, enable learn cluster data, our leverages readily...
Collaborative representation classification (CRC) is an important sparse method, which easy to carry out and uses a linear combination of training samples represent test sample. CRC method utilizes the offset between result each class sample implement classification. However, usually cannot well express difference every In this paper, we propose novel for image recognition address above problem. This not only fuses improve accuracy recognition, but also has fusion mechanism classify images....
Abstract Purpose Dual respiratory–cardiac gating reduces respiratory and cardiac motion blur in myocardial perfusion single‐photon emission computed tomography (MP‐SPECT). However, image noise is increased as detected counts are reduced each dual gate (DG). We aim to develop a denoising method for MP‐SPECT images using 3D conditional generative adversarial network (cGAN). Methods Twenty extended cardiac‐torso phantoms with various 99m Tc‐sestamibi distributions, defect characteristics, body...
Histopathological images contain morphological markers of disease progression that have diagnostic and predictive values, with many computer-aided diagnosis systems using common deep learning methods been proposed to save time labour. Even though are an end-to-end method, they perform exceptionally well given a large dataset often show relatively inferior results for small dataset. In contrast, traditional feature extraction greater robustness small/medium Moreover, texture...
The development of digital pathology offers a significant opportunity to evaluate and analyze the whole slides disease tissue effectively. In particular, segmentation nuclei from histopathology images plays an important role in quantitatively measuring evaluating acquired diseased tissue. There are many automatic methods segment cell images. One widely used unsupervised approach is based on standard k-means or fuzzy c-means (FCM) process color nuclei. Compared with supervised learning...
Burn wound depth is a significant determinant of patient treatment. Typically, the evaluation burn relies heavily on clinical experience doctors. Even experienced surgeons may not achieve high accuracy and speed in diagnosing depth. Thus, intelligent classification useful valuable. Here, an method for based machine learning techniques proposed. In particular, this involves extracting color, texture, features from images, sequentially cascading these features. Then, iterative selection random...
Recent developments in statistical modeling of various linguistic phenomena have shown that additional features give consistent performance improvements. Quite often, improvements are limited by the number a system is able to explore. This paper describes novel progressive training algorithm selects from virtually unlimited feature spaces for conditional maximum entropy (CME) modeling. Experimental results edit region identification demonstrate benefits selection (PFS) algorithm: PFS...
We consider the problem of semi-supervised graph-based learning upon multimodal and mixmodal data. Since in settings, labeled information is very limited, we first propose a non-convex sparse-coding based label propagation (αα-SLP) method to estimate soft labels, thereby enrich supervised information. By considering structural properties data, present embedding (SGE) approach that incorporates with hierarchical local geometric within-class, between-class, overall-class Based on this,...
Plant leaf species classification is an active research area at present with many scientists attempting to use different classifiers features solve it. In this paper we evaluate 10 common classifiers: k-Nearest Neighbors (KNN), support vector machine (SVM), nu-SVM, decision tree, random forest, naïve bayes, linear discriminant analysis (LDA), logistic regression, quadratic (QDA) and sparse representation in such as shape, texture margin. Besides this, numbers of training samples for were...
The two-dimensional principal component analysis (2D-PCA) method has been widely applied in fields of image classification, computer vision, signal processing and pattern recognition. 2D-PCA algorithm also a satisfactory performance both theoretical research real-world applications. It not only retains main information the original face images, but decreases dimension images. In this paper, we integrate spare representation classification (SRC) to distinguish which great novel obtained using...
We consider the problem of semi-supervised graph-based learning. Since in settings, labeled information is limited, we first propose l <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">α</sub> -norm-based label propagation (α-SLP) model to estimate soft labels by using small set and large amount unlabeled training data, thereby enrich supervised information. Based on a-SLP results, conduct discriminant analysis present embedding (SGE) approach...
Inserting an object into a background scene has wide applications in image editing and mixed reality. However, existing methods still struggle to seamlessly adapt the while maintaining its individual characteristics. In this paper, we propose fine-tune pre-trained diffusionbased insertion model such that it learns establish unique correspondence between few weights target object, given as input few-shot images of object. A novel individualized feature extraction (IFE) module is designed...
Pre-trained text-to-image diffusion models are increasingly applied to real-world image super-resolution (Real-ISR) task. Given the iterative refinement nature of models, most existing approaches computationally expensive. While methods such as SinSR and OSEDiff have emerged condense inference steps via distillation, their performance in restoration or details recovery is not satisfied. To address this, we propose TSD-SR, a novel distillation framework specifically designed for...
Segmenting internal structure from echocardiography is essential for the diagnosis and treatment of various heart diseases. Semi-supervised learning shows its ability in alleviating annotations scarcity. While existing semi-supervised methods have been successful image segmentation across medical imaging modalities, few attempted to design specifically addressing challenges posed by poor contrast, blurred edge details noise echocardiography. These characteristics pose generation high-quality...
Owing to the robust priors of diffusion models, recent approaches have shown promise in addressing real-world super-resolution (Real-SR). However, achieving semantic consistency and perceptual naturalness meet human perception demands remains difficult, especially under conditions heavy degradation varied input complexities. To tackle this, we propose Hero-SR, a one-step diffusion-based SR framework explicitly designed with priors. Hero-SR consists two novel modules: Dynamic Time-Step Module...