A5072609108- Advanced MRI Techniques and Applications
- Image and Signal Denoising Methods
- Medical Imaging Techniques and Applications
- Medical Image Segmentation Techniques
- Advanced Image Fusion Techniques
- Photoacoustic and Ultrasonic Imaging
- Radiomics and Machine Learning in Medical Imaging
- Advanced Image Processing Techniques
- Advanced Vision and Imaging
- AI in cancer detection
- Brain Tumor Detection and Classification
- Research on Leishmaniasis Studies
- Chemistry and Chemical Engineering
- Multiple Myeloma Research and Treatments
- Image Enhancement Techniques
- Optical measurement and interference techniques
- Sparse and Compressive Sensing Techniques
- Image Retrieval and Classification Techniques
- Remote-Sensing Image Classification
- Image Processing Techniques and Applications
- Computer Graphics and Visualization Techniques
- Advanced Neural Network Applications
- Infrared Thermography in Medicine
- COVID-19 diagnosis using AI
- Machine Learning in Materials Science
University of South China
2018-2024
South China University of Technology
2023
Huazhong University of Science and Technology
2015-2018
Multimodal medical image fusion aims to integrate complementary information from different modalities of images. Deep learning methods, especially recent vision Transformers, have effectively improved performance. However, there are limitations for Transformers in fusion, such as lacks local feature extraction and cross-modal interaction, resulting insufficient multimodal integration. In addition, the computational cost is higher. To address these challenges, this work, we develop an...
It is crucial to integrate the complementary information of multimodal medical images for enhancing image quality in clinical diagnosis. Convolutional neural network (CNN) based deep learning methods have been widely utilized fusion due their strong modeling ability. However, CNNs fail build long-range dependencies an image, which limits performance. To address this issue, work, we develop a new unsupervised framework that combines Swin Transformer and CNN. The proposed model follows...
This paper intends to generate the approximate Voronoi diagram in geodesic metric for some unbiased samples selected from original points. The mesh model of seeds is then constructed on basis diagram. Rather than constructing all points, proposed strategy run around obstacle that distances among neighboring points are sensitive nearest neighbor definition. It obvious reconstructed level detail Hence, our main motivation deal with redundant scattered In implementation, Poisson disk sampling...
Abstract Magnetic resonance imaging (MRI) is a non‐interposition technique that provides rich anatomical and physiological information. Yet it limited by the long time. Recently, deep neural networks have shown potential to significantly accelerate MRI. However, most of these approaches ignore correlation between adjacent slices in MRI image sequences. In addition, existing learning‐based methods for are mainly based on convolutional (CNNs). They fail capture long‐distance dependencies due...
As an advanced medical imaging technology, magnetic resonance (MRI) has great advantages and application potentials in clinical diagnosis. However, since the long scanning time artifacts caused by patient movements, results are always not satisfactory. Therefore, accelerating MRI improving quality key problems. In this work, we propose a novel deep network that combines U-net architecture with non-local attention blocks for reconstruction. We employ to construct basic network. The is...
Drug combination therapy is a powerful solution for the treatment of complex disease such as cancers due to its capability therapeutic efficacy and reducing side effects. Nevertheless, it very difficult screen all drug combinations by experiments since vast number possible combinations. Currently, computational methods, especially graph neural networks transformer, have been developed discover prioritization shown promising potentials. Despite great achievements obtained existing models,...
Multimodal medical image fusion is vital for extracting complementary information and generating comprehensive images in clinical applications. However, existing deep learning-based approaches face challenges effectively utilizing frequency-domain information, designing appropriate integration strategies modelling long-range context correlation. To address these issues, we propose a novel unsupervised multimodal method called Multiscale Fourier Attention Detail-Aware Fusion (MFA-DAF). Our...
Multi-contrast magnetic resonance (MR) images are crucial for diagnosing diseases and analysis in clinical practice, yet their acquisition often entails long scanning procedures. Recently, exploiting shared information among multi-contrast MR yields favorable results reconstruction. However, most studies simply concatenate the without effective matching fusion mechanisms to mitigate impact of redundant features. Furthermore, feature extraction backbone networks tend lose high-frequency...
Magnetic resonance imaging (MRI) is an technology widely used in medical clinical diagnosis. Nevertheless, it always takes a long time to obtain the high spatial resolution MR image. Image reconstruction has been playing crucial role for accelerated MRI. Recently, deep neural networks show potential significantly speed up However, these learning methods usually learn feature information domain. Although some works have attempted utilize frequency from k-space reconstruct images, they still...
Two-stage image inpainting method based on transformer and Fast Fourier convolution
In clinical applications, medical image reconstruction is crucial for extracting complementary information and restoring quality. However, existing deep learning-based methods suffer from the following two problems: first, they ignore frequency-domain specific to magnetic resonance imaging, second, most of traditional networks are single-end-to-end networks, which lack effective attention feature information. We suggest a unique two-stage progressive network (TSPNet) solve these issues. Our...
In recent years, liver cancer has become one of the five dangerous cancers due to highest mortality ratios worldwide. Automatic tumor segmentation is a most important task help radiologists and oncologists analyze CT images. With rapid development Convolutional Neural Network (CNN), UNet2D have been widely applied in medical image segmentation. But 2D convolutions cannot extract more spatial information, making it difficult for network learn powerful features between slices. order address...