A5102819028- Advanced Neural Network Applications
- Medical Image Segmentation Techniques
- Radiomics and Machine Learning in Medical Imaging
- COVID-19 diagnosis using AI
- Medical Imaging and Analysis
- AI in cancer detection
- Brain Tumor Detection and Classification
- Hepatocellular Carcinoma Treatment and Prognosis
- Advanced X-ray and CT Imaging
- Pancreatic and Hepatic Oncology Research
- Medical Imaging Techniques and Applications
- Domain Adaptation and Few-Shot Learning
- MRI in cancer diagnosis
- Power Systems Fault Detection
- Advanced MRI Techniques and Applications
- Ultrasonics and Acoustic Wave Propagation
- Venous Thromboembolism Diagnosis and Management
- Islanding Detection in Power Systems
- Colorectal Cancer Screening and Detection
- Power Quality and Harmonics
- Heart Rate Variability and Autonomic Control
- Image Retrieval and Classification Techniques
- Frailty in Older Adults
- HVDC Systems and Fault Protection
- Anomaly Detection Techniques and Applications
Zhejiang University
2015-2025
Zhejiang Lab
2020-2024
The detection and delineation of the liver from abdominal 3D computed tomography (CT) images are fundamental tasks in computer-assisted surgery planning. However, automatic accurate segmentation, especially detection, remains challenging due to complex backgrounds, ambiguous boundaries, heterogeneous appearances highly varied shapes liver. To address these difficulties, we propose an segmentation framework based on convolutional neural network (CNN) globally optimized surface evolution....
This paper proposes an adaptive protection scheme for distribution systems with distributed generations (DGs). To mitigate the influences on devices by DGs, online calculation method of fault current under various system operation conditions is presented. By sampling local available measurements in buses, a proposed optimized estimation used to calculate Thevenin equivalent circuit parameters dynamically. Moreover, short-circuit can be calculated accurately set overcurrent relay value...
Efficient and accurate 3D liver segmentations from contrast-enhanced computed tomography (CT) images play an important role in therapeutic strategies for hepatic diseases. However, inhomogeneous appearances, ambiguous boundaries, large variance shape often make it a challenging task. The existence of abnormalities poses further difficulty. Despite the significant intensity difference, tumors should be segmented as part liver. This study aims to address these challenges, especially when...
Pancreas identification and segmentation is an essential task in the diagnosis prognosis of pancreas disease. Although deep neural networks have been widely applied abdominal organ segmentation, it still challenging for small organs (e.g. pancreas) that present low contrast, highly flexible anatomical structure relatively region. In recent years, coarse-to-fine methods improved accuracy by using coarse predictions fine stage, but only object location utilized rich image context neglected....
Abstract In the service of composite materials, scenarios involving repeated impacts are frequently encountered. our study, time‐frequency characteristics acoustic emission signals during impact processes were investigated using wavelet packet transform. Principal Component Analysis was utilized to identify frequency bands containing most information. Shannon entropy employed select optimal basis function. The terms DDD3, ADD3, and DAA3 correspond matrix cracking, delamination, fiber...
The pancreas plays an important role in glucose metabolism, and developing diabetes or long-term metabolism disturbance may be a prevalent sequela after pancreatectomy. Nevertheless, relative factors of new-onset pancreatectomy stay unclear. Radiomics analysis is potential to identify image markers for disease prediction prognosis. Meanwhile, combination imaging electronic medical record (EMR) showed superior performance than EMR alone previous studies. One critical step identity predictors...
Accurate pancreas and pancreatic tumor segmentation from abdominal scans is crucial for diagnosing treating diseases. Automated reliable algorithms are highly desirable in both clinical practice research.
This paper presents a new version of Dropout called Split (sDropout) and rotational convolution techniques to improve CNNs' performance on image classification. The widely used standard has advantage preventing deep neural networks from overfitting by randomly dropping units during training. Our sDropout splits the data into two subsets keeps both rather than discards one subset. We also introduce techniques, i.e. rotate-pooling (RPC) flip-rotate-pooling (FRPC) boost robustness for rotation...
Purpose Segmentation of the liver from abdominal computed tomography (CT) image is an essential step in some computer assisted clinical interventions, such as surgery planning for living donor transplant (LDLT), radiotherapy and volume measurement. In this work, we develop a deep learning algorithm with graph cut refinement to automatically segment CT scans. Methods The proposed method consists two main steps: (i) simultaneously detection probabilistic segmentation using 3D convolutional...
Segmentation of pancreatic tumors on CT images is essential for the diagnosis and treatment cancer. However, low contrast between pancreas tumor, as well variable tumor shape position, makes segmentation challenging. To solve problem, we propose a Position Prior Attention Network (PPANet) with pseudo generation module (PSGM) position prior attention (PPAM). PSGM PPAM maps to latent space generate map supervises location classification. The proposed method evaluated patient data collected...
Computed tomography (CT) is widely applied in contemporary clinic. Due to the radiation risks carried by X-rays, imaging and post-processing methods of low-dose CT (LDCT) become popular topics academia industrial community. Generally, LDCT presents strong noise artifacts, while existing algorithms cannot completely overcome blurring effects meantime reduce noise. The proposed method enables extend independent frequency channels wavelet transformation, then two separate networks are...
Pancreatic cancer is a highly malignant of the digestive tract and rapidly progressing spreading clinically. Automatic accurate pancreatic tissue segmentation in abdominal CT images essential for early diagnosis pancreatic-related diseases. It challenging that pancreas small size complex morphology. To address this problem, we propose dual-attention model fusing CNN Transformer to effectively activate pancreas-related features expression. The structure weights importance at channel level...
Pulmonary embolism (PE) is an important clinical disorder that will result in lung tissue damage or low blood oxygen levels, which need early diagnosis and timely treatment. While computed tomographic pulmonary angiography (CTPA) the gold standard to diagnose PE, previous studies have verified effectiveness of combing CTPA EMR data computer-aided PE detection diagnosis. In this paper, we proposed a multimodality fusion method based on multi-view subspace clustering guided feature selection...
Survival prediction is crucial for treatment decision making in hepatocellular carcinoma (HCC). We aimed to build a fully automated artificial intelligence system (FAIS) that mines whole-liver information predict overall survival of HCC. included 215 patients with preoperative contrast-enhance CT imaging and received curative resection from hospital China. The cohort was randomly split into developing testing subcohorts. FAIS constructed convolutional layers full-connected layers. Cox...
With the development of radiology and computer technology, diagnosis by medical imaging is heading toward precision automation. Due to complex anatomy around pancreatic tissue high demands for clinical experience, assisted pancreas segmentation system will greatly promote efficiency. However, existing model suffers from poor generalization among images multiple hospitals. In this paper, we propose an end-to-end data-adaptive tackle problems lack annotations generalizability. The employs...
Abstract To study the effect of short cardio on obese male college students’ inhibitory control ability, we investigate difference brain activation patterns before and after by analyzing resting state fMRI data. The experiment results this show that can improve ability people.
Computer-aided automatic pancreas segmentation is essential for early diagnosis and treatment of pancreatic diseases. However, the annotation images requires professional doctors considerable expenditure. Due to imaging differences among various institution population, scanning devices, protocols, so on, significant degradation in performance model inference results prone occur when models trained with domain-specific (usually institution-specific) datasets are directly applied new (other...
Purpose: MRI guided radiotherapy is particularly attractive for abdominal targets with low CT contrast. To fully utilize this modality pancreas tracking, automated segmentation tools are needed. A hybrid gradient, region growth and shape constraint (hGReS) method to segment 2D upper dynamic developed purpose. Methods: coronal MR images of 2 healthy volunteers were acquired a frame rate 5 f/second. The regions interest (ROIs) included the liver, stomach. first was used as source where centers...