A5100423724- Topic Modeling
- Natural Language Processing Techniques
- Pancreatic and Hepatic Oncology Research
- Software Engineering Research
- Software Testing and Debugging Techniques
- Cancer-related cognitive impairment studies
- Magnetic properties of thin films
- Cancer survivorship and care
- Magnetic Field Sensors Techniques
- Speech and dialogue systems
- Magnetic Properties and Applications
- Face and Expression Recognition
- Biometric Identification and Security
- Image Retrieval and Classification Techniques
- Forensic Fingerprint Detection Methods
- Service-Oriented Architecture and Web Services
- Software Reliability and Analysis Research
- Software System Performance and Reliability
- Quantum Computing Algorithms and Architecture
- Advanced Image Processing Techniques
- Machine Learning and Data Classification
- Safety and Risk Management
- Perovskite Materials and Applications
- Metallurgy and Cultural Artifacts
- Thyroid Cancer Diagnosis and Treatment
Chinese Academy of Medical Sciences & Peking Union Medical College
2024-2025
Fuzhou University
2025
Tan Kah Kee Innovation Laboratory
2025
Shanghai Medical College of Fudan University
2023-2024
Shanghai Cancer Institute
2023-2024
Fudan University Shanghai Cancer Center
2023-2024
Nantong University
2024
Nankai University
2009-2024
Fudan University
2023-2024
Central South University
2024
Abstract Background There is no evidence supporting the feasibility of laparoscopic pancreaticoduodenectomy (LPD) compared to open pancreatoduodenectomy (OPD) following neoadjuvant chemotherapy (NACT) for pancreatic ductal adenocarcinoma (PDAC). Methods The clinical data consecutive patients with borderline resectable PDAC who received NACT and underwent either LPD or OPD between January 2020 December 2022 at Fudan University Shanghai Cancer Center was prospectively collected retrospectively...
Large language models (LLMs), such as ChatGPT, have received substantial attention due to their capabilities for understanding and generating human language. While there has been a burgeoning trend in research focusing on the employment of LLMs supporting different medical tasks (e.g., enhancing clinical diagnostics providing education), review these efforts, particularly development, practical applications, outcomes medicine, remains scarce. Therefore, this aims provide detailed overview...
Abstract Large ZnSe quantum dots (QDs) with an emission peak ≈450 nm hold significant promise for display technologies. However, achieving efficient pure‐blue through the enlargement of nanocrystals remains a challenge. In this study, breakthrough is reported in growing large‐size QDs well beyond exciton Bohr radius Yb 3+ doping strategy. reduces surface energy (220) crystal plane and alleviates interface strain ZnSe/ZnS structure, enabling to grow larger while maintaining enhanced...
Thermal ablation (TA) modalities such as radiofrequency (RFA), microwave (MWA), and laser (LA) have been widely used in the treatment of papillary thyroid microcarcinoma (PTMC) patients. Based on previous small-sample meta-analyses, this network meta was designed to further compare efficacy safety these thermal methods PTMC
Human Activity Recognition (HAR) has recently attracted widespread attention, with the effective application of this technology helping people in areas such as healthcare, smart homes, and gait analysis. Deep learning methods have shown remarkable performance HAR. A pivotal challenge is trade-off between recognition accuracy computational efficiency, especially resource-constrained mobile devices. This necessitates development models that enhance feature representation capabilities without...
Objective. Intestinal metaplasia (IM) is a common precancerous condition for gastric cancer, and the risk of developing cancer increases as IM worsens. However, current deep learning-based methods cannot effectively model complex geometric structure lesions. To accurately diagnose severity prevent occurrence we revisit deformable convolution network (DCN), propose novel offset generation method based on color features to guide convolution, named color-guided convolutional (CDCN).Approach....
The performance of automatic fingerprint identification system relies heavily on the quality captured images. A novel method for image analysis has been presented, which overcomes shortcoming most existing methods have, considering correlation each feature as linear and paying no attention to clarity local texture. In this paper, ten features are extracted from then Fuzzy Relation Classifier is trained classify images, includes unsupervised clustering supervised classification care more...
Three-dimensional video coding is one of the main challenges restricting widespread applications 3D and free viewpoint video.In this paper, a novel fractal algorithm with motion-vector-field-based motion estimation for depth map sequence proposed.We firstly add pre-search restriction to rule improper domain blocks out matching search process so that number involved in can be restricted smaller size.Some improvements including initial point prediction, threshold transition condition early...
A bstract In 2020, Novartis Pharmaceuticals Corporation and the U.S. Food Drug Administration (FDA) started a 4-year scientific collaboration to find novel radiogenomics-based prognostic predictive factors for HR+/HER2-metastatic breast cancer under Research Collaboration Agreement. This manuscript aims detail guiding principles methodology this study. We include discussion of internal external clinical, genomics, imaging datasets, data processing workflows, machine learning model...
The goal of session-based recommendation in E-commerce is to predict the next item that an anonymous user will purchase based on browsing and history. However, constructing global or local transition graphs supplement session data can lead noisy correlations intent vanishing. In this work, we propose Frequent Attribute Pattern Augmented Transformer (FAPAT) characterizes intents by building attribute matching patterns. Specifically, frequent compact patterns are served as memory augment...
Fingerprint image enhancement is very important in AFIS (Automatic Identification System). In this paper, a fingerprint method based on Gabor filtering introduced. Compared with traditional Gabor-based method, the algorithm uses improved to improve clarity of ridge structures fingerprints. The can protect structure and avoid spurious ridges effectively. addition, Fourier-Transformation proposed compute orientation field more precisely efficiently. Experimental results show that not only...
Abstract Causality is an important element in decision-making and interventions are required to optimize results of target values. In this paper, based on the model Causal Bayesian Optimization, a counter-noise version acquisition function proposed new prior estimation algorithms including Support Vector Regression, Ridge Regression Random Forest evaluated. This paper provides improved framework facilitate causal inference optimization processes.
Tabular data, as a crucial form of data representation, exists in diverse formats on the Web. When confronted with complex and irregular tables, manual modification becomes laborious task. This paper investigates performance Large Language Models (LLMs) context table editing tasks. Existing research mainly focuses regular-shaped wherein instructions are used to generate code SQL, Python, or Excel Office-script for manipulating tables. Nevertheless, tables structures, particularly those...