A5100419510- Radiomics and Machine Learning in Medical Imaging
- Advanced Neural Network Applications
- AI in cancer detection
- Advanced Image and Video Retrieval Techniques
- Medical Image Segmentation Techniques
- Image Retrieval and Classification Techniques
- Medical Imaging Techniques and Applications
- Remote-Sensing Image Classification
- Formal Methods in Verification
- Neural Networks and Applications
- Functional Brain Connectivity Studies
- Advanced Algorithms and Applications
- Brain Tumor Detection and Classification
- Image and Signal Denoising Methods
- Remote Sensing and Land Use
- Advanced Vision and Imaging
- Breast Cancer Treatment Studies
- Advanced MRI Techniques and Applications
- Digital Radiography and Breast Imaging
- Breast Lesions and Carcinomas
- Video Surveillance and Tracking Methods
- Anomaly Detection Techniques and Applications
- Advanced Computational Techniques and Applications
- Advanced Neuroimaging Techniques and Applications
- Multimodal Machine Learning Applications
University of Electronic Science and Technology of China
2016-2024
Sun Yat-sen University Cancer Center
2018-2024
Sun Yat-sen University
2018-2024
Xi'an University of Science and Technology
2021-2024
Management and Science University
2024
Sichuan Agricultural University
2021-2023
Taiyuan University of Technology
2023
Qilu University of Technology
2023
Chengdu Third People's Hospital
2023
Southwest Jiaotong University
2023
Intrinsic neural activity ubiquitously persists in all physiological states. However, how intrinsic brain (iBA) changes over a short time remains unknown. To uncover the dynamics' theoretic underpinning, electrophysiological relevance, and neuromodulation, we identified iBA dynamics on simulated data, electroencephalogram-functional magnetic resonance imaging (EEG-fMRI) repetitive transcranial stimulation (rTMS) fMRI data using sliding-window analysis. The temporal variability (dynamics) of...
Mammography is successfully used as an effective screening tool for cancer diagnosis. A calcification cluster on mammography a primary sign of cancer. Early researches have proved the diagnostic value calcification, yet their performance highly dependent handcrafted image descriptors. Characterizing in automatic and robust way remains challenge. In this paper, was characterized by descriptors obtained from deep learning We compared performances different feature sets digital mammograms. The...
The traditional de novo drug discovery is known as a high cost and risk process. In response, recently there an increasing interest in discovering new indications for drugs-a process repositioning-using computational methods. this study, we present systematic approach identifying potential of existing through its relation to similar drugs. Different from the previous similarity-based methods, adapted novel bipartite-graph based method when considering common targets their interaction...
Corona Virus Disease (COVID-19) has spread globally quickly, and resulted in a large number of causalities medical resources insufficiency many countries. Reverse-transcriptase polymerase chain reaction (RT-PCR) testing is adopted as biopsy tool for confirmation virus infection. However, its accuracy low 60-70%, which inefficient to uncover the infected. In comparison, chest CT been considered prior choice diagnosis monitoring progress COVID-19 Although diagnostic systems based on artificial...
Abstract Breast carcinoma is the second largest cancer in world among women. Early detection of breast has been shown to increase survival rate, thereby significantly increasing patients’ lifespan. Mammography, a noninvasive imaging tool with low cost, widely used diagnose disease at an early stage due its high sensitivity. Although some public mammography datasets are useful, there still lack open access that expand beyond white population as well missing biopsy confirmation or unknown...
The accuracy of fish farming and real-time monitoring are essential to the development "intelligent" farming. Although existing instance segmentation networks (such as Maskrcnn) can detect segment fish, most them not effective in monitoring. In order improve image promote accurate intelligent industry, this article uses YOLOv5 backbone network object detection branch, combined with semantic head for segmentation. experiments show that precision reach 95.4% 98.5% algorithm structure proposed...
Bacterial vaginosis (BV) is caused by the excessive and imbalanced growth of bacteria in vagina, affecting 30 to 50% women. Gram staining followed Nugent scoring based on bacterial morphotypes under microscope considered gold standard for BV diagnosis; this method often labor-intensive time-consuming, results vary from person person. We developed optimized a convolutional neural network (CNN) model evaluated its ability automatically identify classify three categories scores images.
Abstract Poststroke aphasia (PSA) results from direct effect of focal lesions and dysfunction distributed language networks. However, how flexible the activity at specific nodes control global dynamics is currently unknown. In this study, we demonstrate that alterations in regional may cause imbalances between segregation integration temporo‐spatial pattern, transient are disrupted PSA patients. Specifically, applied dynamic framework to eyes‐closed resting‐state functional MRI data patients...
Abstract Several neuroimaging studies have examined cerebral function in patients who suffer from aphasia, but the mechanism underlying this disorder remains poorly understood. In study, we alterations local regional and remote interregional network functions aphasia combined with amplitude of low-frequency fluctuations functional connectivity (FC) using resting-state magnetic resonance imaging. A total 17 post-stroke aphasic patients, all having suffered a stroke left hemisphere, as well 20...
Breast lesion is a malignant tumor that occurs in the epithelial tissue of breast. The early detection breast lesions can make patients for treatment and improve survival rate. Thus, accurate automatic segmentation from ultrasound images fundamental task. However, effectively still faced up with two challenges. One characteristics lesions’ multi-scale other one blurred edges difficult. To solve these problems, we propose deep learning architecture, named Multi-scale Fusion U-Net (MF U-Net),...
Micro-expressions (MEs) spotting is popular in some fields, for example, criminal investigation and business communication. But it still a challenging task to spot the onset offset of MEs accurately long videos. This paper refines every step workflow before feature extraction, which can reduce error propagation. The takes advantage high-quality alignment method, more accurate landmark detector, also robust optical flow estimation. Besides, Bayesian optimization hybrid with Nash equilibrium...
Image-recipe retrieval, which aims at retrieving the relevant recipe from a food image and vice versa, is now attracting widespread attention, since sharing food-related images recipes on Internet has become popular trend. Existing methods have formulated this problem as typical cross-modal retrieval task by learning image-recipe similarity. Though these made inspiring achievements for they may still be less effective to jointly incorporate three crucial points: (1) association between...
In recent years, the Internet has stimulated explosion of multimedia data. Food-related cooking videos, images, and recipes promote rapid development food computing. Image-recipe retrieval is an important sub-task in field cross-modal retrieval, which focuses on measurement association between image recipe (title, ingredients, instructions). Although existing methods have proposed some feasible solutions to achieve goal there are still following issues: 1) complex model structure...
The current target tracking and detection algorithms often have mistakes omissions when the is occluded or small. To overcome defects, this paper integrates bi-directional feature pyramid network (BiFPN) into cascade region-based convolutional neural (R-CNN) for live object detection. Specifically, BiFPN structure was utilized to connect between scales fuse weighted features more efficiently, thereby enhancing network’s extraction ability, improving effect on small targets. proposed method,...
To evaluate whether contrast-enhanced cone-beam breast CT (CE-CBBCT) features can risk-stratify prognostic stage in cancer.Overall, 168 biopsy-proven cancer patients were analysed: 115 the training set underwent scanning using v. 1.5 CE-CBBCT between August 2019 and December 2019, whereas 53 test 1.0 May 2012 2014. All restaged according to American Joint Committee on Cancer eighth edition staging system. Following combination of imaging parameters clinicopathological factors, predictors...
Abstract Road crack detection is an important task for road safety and maintenance. In the past, people made use of manual methods tried to computer vision detect crack. The most prominent feature in recent years deep learning. However, there no good learning method under noise. This challenge faced bravely. First, a noise dataset proposed, consisting multiple images which called NCD. Then, adaptive bilateral filtering algorithm developed, can reduce influence Finally, new network with two...
Computer-aided diagnosis has emerged as a rapidly evolving field, garnering increased attention in recent years. At the forefront of this field is segmentation lesions medical images, which critical preliminary stage subsequent treatment procedures. Among most challenging tasks image analysis accurate and automated brain tumors various modalities tumor MRI. In article, we present novel end-to-end network architecture called MMGan, combines advantages residual learning generative adversarial...
Brain tumors are the brain diseases with highest mortality and prevalence, magnetic resonance imaging has high-resolution multiparameter. As basis for realizing quantitative analysis of tumors, automatic segmentation plays a vital role in diagnosis treatment. A new network model is proposed to improve accuracy convolutional neural tumor regions control parameter space scale model. The first uses layer composed series 3D convolution filters construct backbone feature learning input MRI image...