A5003171599- AI in cancer detection
- Digital Imaging for Blood Diseases
- Cell Image Analysis Techniques
- Remote-Sensing Image Classification
- Radiomics and Machine Learning in Medical Imaging
- Advanced Image Processing Techniques
- Pancreatic and Hepatic Oncology Research
- Neuroendocrine Tumor Research Advances
- Gene expression and cancer classification
- Semiconductor materials and devices
- Digital Media Forensic Detection
- COVID-19 diagnosis using AI
- Ferroelectric and Negative Capacitance Devices
- MXene and MAX Phase Materials
- Image and Signal Denoising Methods
East China Normal University
2022-2024
Endoscopic ultrasonography-guided fine-needle aspiration/biopsy (EUS-FNA/B) is considered to be a first-line procedure for the pathological diagnosis of pancreatic cancer owing its high accuracy and low complication rate. The number new cases ductal adenocarcinoma (PDAC) increasing, accurate poses challenge cytopathologists. Our aim was develop hyperspectral imaging (HSI)-based convolution neural network (CNN) algorithm aid in EUS-FNA cytology specimens.
The hafnia-based ferroelectric oxides with excellent negative-capacitance properties offer a great opportunity to develop high-performance integrated circuits. nanosized multiphase distribution of Hf0.5Zr0.5O2 (HZO) has significant influence on its properties. Transmission electron microscope (TEM) an atomistic resolution could establish the structure-property relationship and guide performance improvement HZO by identifying phase structures. However, high throughput TEM data complexity...
RGB images and medical hyperspectral (MHSIs) are two widely-used modalities in computational pathology. The former is cheap, easy fast to obtain while lacking pathological information such as physiochemical state. latter an emerging modality which captures electromagnetic radiation matter interaction but suffers from problems high time cost low spatial resolution. In this paper, we bring forward a unified dual-task multi-modality self-supervised learning (SSL) framework, called Uni-Dual,...
Hyperspectral images (HSIs) offer great potential for computational pathology. However, limited by the spectral redundancy and lack of prior in popular 2D networks, previous HSI based techniques do not perform well. To address these problems, we propose to segment HSIs from a deformable perspective, which processes different bands independently fuses spatiospectral features interest via attention mechanisms. In addition, Deformable Self-Supervised Spectral Regression (DF-S3R), introduces two...
Histology analysis of the tumor micro-environment integrated with genomic assays is gold standard for most cancers in modern medicine. This paper proposes a Gene-induced Multimodal Pre-training (GiMP) framework, which jointly incorporates genomics and Whole Slide Images (WSIs) classification tasks. Our work aims at dealing main challenges multi-modality image-omic w.r.t. (1) patient-level feature extraction difficulties from gigapixel WSIs tens thousands genes, (2) effective fusion...
Benefited from the rich and detailed spectral information in hyperspectral images (HSI), HSI offers great potential for a wide variety of medical applications such as computational pathology. But, lack adequate annotated data high spatiospectral dimensions HSIs usually make classification networks prone to overfit. Thus, learning general representation which can be transferred downstream tasks is imperative. To our knowledge, no appropriate self-supervised pre-training method has been...