A5100704700- Anomaly Detection Techniques and Applications
- Human Pose and Action Recognition
- Video Surveillance and Tracking Methods
- Advanced Vision and Imaging
- Advanced Image Processing Techniques
- Image Processing Techniques and Applications
- Video Coding and Compression Technologies
- Advanced Wireless Communication Technologies
- Advanced MIMO Systems Optimization
- Vascular Tumors and Angiosarcomas
- Energy Harvesting in Wireless Networks
- Artificial Intelligence in Law
- Thyroid and Parathyroid Surgery
- Semiconductor materials and devices
- Recommender Systems and Techniques
- Cardiac tumors and thrombi
- Advancements in Semiconductor Devices and Circuit Design
- Topic Modeling
- Medical Imaging Techniques and Applications
- Sarcoma Diagnosis and Treatment
- Ferroelectric and Negative Capacitance Devices
- Privacy-Preserving Technologies in Data
- Orthopaedic implants and arthroplasty
- Advanced Neural Network Applications
- Gait Recognition and Analysis
Shanghai Jiao Tong University
2017-2025
Shandong University
2008-2024
Qingdao Binhai University
2023-2024
University of Electronic Science and Technology of China
2015-2024
Cornell University
2024
Weill Cornell Medicine
2024
Stony Brook University
2023-2024
South China Normal University
2016-2024
Jiangxi Normal University
2016-2023
Second Affiliated Hospital of Zhejiang University
2021-2023
Real-time positioning feature is becoming an essential part in a variety of military and civilian applications. In particular, the cooperation multisensors helps to improve navigation accuracy universality system. this article, we propose novel robust fault-tolerant federated filter based on strapdown inertial system (SINS), global satellite (GNSS), celestial (CNS), doppler velocity log (DVL). proposed algorithm, position information GNSS, DVL, attitude CNS are introduced as measurement...
Cyber threat intelligence (CTI) sharing has gradually become an important means of dealing with security threats. Considering the growth cyber intelligence, quick analysis threats a hot topic at present. Researchers have proposed some machine learning and deep models to automatically analyze these immense amounts intelligence. However, due large amount network terminology in CTI, based on open-domain corpus perform poorly CTI automatic task. To address this problem, we propose method named...
The Li dendrite growth during battery cycles is a well‐known obstacle to the practical application of sulfide‐based electrolytes (SEs), notably Li3PS4 (LPS), in lithium metal batteries. However, there remains significant gap understanding mechanism for dendrites penetration through SEs exhibiting high shear modulus. Herein, we investigate optimum deposition sites Li0 atoms within typical LPS configurations, encompassing crystalline, lithiated, and degraded structures, with their ionization...
The Li dendrite growth during battery cycles is a well‐known obstacle to the practical application of sulfide‐based electrolytes (SEs), notably Li3PS4 (LPS), in lithium metal batteries. However, there remains significant gap understanding mechanism for dendrites penetration through SEs exhibiting high shear modulus. Herein, we investigate optimum deposition sites Li0 atoms within typical LPS configurations, encompassing crystalline, lithiated, and degraded structures, with their ionization...
Cardiac myxomas are the most common primary neoplasms of heart. The present study was performed on 61 cases patients with cardiac myxoma, in order to investigate tumors' clinical and pathological features, identify relationship between characteristics behaviors.A total myxoma were analyzed reviewed retrospectively, including presentations, physical examinations, echocardiography, electrocardiography, pathology documents.The patient cohort made up 37 women 24 men. average age at diagnosis...
Magnetic resonance imaging (MRI) vertebral localization, identification, and segmentation are important steps in the automatic analysis of spines. Due to similar appearances vertebrae, accurate segmentation, identification vertebrae remain challenging. Previous methods solved three tasks independently, ignoring intrinsic correlation among them. In this paper, we propose a multi-task relational learning network (MRLN) that utilizes both relationships between relevance tasks. A dilation...
The recognition performance of existing vision-based human action (HAR) methods is greatly reduced in the case low camera resolution or occlusion. Wearable sensors can provide complementary information to alleviate this problem. It challenging construct a robust HAR model using multimodal wearable-sensor data. In paper, we propose cross-Attention-based Multimodal fusion Wavelet Knowledge Distillation Network (MAWKDN) method guide from video data by acquiring wearable and reduce noise effects...
The traditional text classification methods usually follow this process: first, a sentence can be considered as bag of words (BOW), then transformed into feature vector which classified by some methods, such maximum entropy (ME), Naive Bayes (NB), support machines (SVM), and so on. However, when these are applied to classification, we not obtain an ideal result. most important reason is that the semantic relations between very for categorization, however, method capture it. Sentiment special...
Segmentation of the left ventricular (LV) myocardium in 2-D echocardiography is essential for clinical decision making, especially geometry measurement and index computation. However, segmenting a time-consuming process challenging due to fuzzy boundary caused by low image quality. The ground-truth label employed as pixel-level class associations or shape regulation segmentation, which works limit effective feature enhancement echocardiography. We propose training strategy named...
Abstract Background Transoral endoscopic thyroidectomy via vestibular approach (TOETVA) has become increasingly popular in the treatment of papillary thyroid cancer (PTC). This study aimed to describe safety and feasibility total between TOETVA open (OT) approaches for patients with PTC. Methods We retrospectively reviewed 780 consecutive suffering from PTC that had undergone using ( n = 107) OT 673) April 2016 December 2021 at our institute. Afterward, a 101 matched patients' surgical...
Thanks to its powerful ability depict high-resolution anatomical information, magnetic resonance imaging (MRI) has become an essential non-invasive scanning technique in clinical practice. However, excessive acquisition time often leads the degradation of image quality and psychological discomfort among subjects, hindering further popularization. Besides reconstructing images from undersampled protocol itself, multi-contrast MRI protocols bring promising solutions by leveraging additional...
For multi-object tracking (MOT), jointly learning the detector and embedding model (JDE) is one of mainstream solutions. However, an inherent problem in this architecture arises as tasks target detection appearance feature extraction compete with each other. FairMOT, a representative method, attempts to address issue by employing two homogeneous branches, but it overlooks essential difference between these tasks. Upon original network architecture, we propose adaptive dual decoder structure....
We report a case of 42-year-old man, who presented with three-weeks' history dyspnea and chest oppression induced by light activity, had large mass in the pericardium. Computed tomography showed non-homogeneous density lobulated pericardial sac. The tumor underwent completely resection histopathological analysis revealed to be pleomorphic liposarcoma, which was composed predominantly epithelioid cells. patient is still alive 18 months after diagnosis made. rare incidence liposarcoma...
Intravascular ultrasound (IVUS) can provide high-resolution cross-sectional images of coronary arteries, showing detailed information the vascular lumen, tube wall, and athermanous plaques, which is helpful for discovery or identification early atherosclerotic plaques. Multiple parameters extractable from IVUS image help cardiologist analyze pathology assist in disease diagnosis postoperative treatment. Typically, cardiologists manually label intima adventitia image, obtain a limited number...