A5105759669- COVID-19 diagnosis using AI
- Tuberculosis Research and Epidemiology
- Machine Learning in Healthcare
- Machine Learning and ELM
- Domain Adaptation and Few-Shot Learning
- Radiomics and Machine Learning in Medical Imaging
- Machine Learning and Algorithms
- AI in cancer detection
Shanghai Jiao Tong University
2023-2024
Shandong Jiaotong University
2023-2024
Vision-Language Pre-training (VLP) that utilizes the multi-modal information to promote training efficiency and effectiveness, has achieved great success in vision recognition of natural domains shown promise medical imaging diagnosis for Chest X-Rays (CXRs). However, current works mainly pay attention exploration on single dataset CXRs, which locks potential this powerful paradigm larger hybrid multi-source CXRs datasets. We identify although blending samples from diverse sources offers...
Class incremental learning (CIL) aims to incrementally update a trained model with the new classes of samples (plasticity) while retaining previously learned ability (stability). To address most challenging issue in this goal, i.e., catastrophic forgetting, mainstream paradigm is memory-replay CIL, which consolidates old knowledge by replaying small number saved memory. Despite effectiveness, inherent destruction-reconstruction dynamics CIL are an intrinsic limitation: if severely...
Many real-world image recognition problems, such as diagnostic medical imaging exams, are "long-tailed" $\unicode{x2013}$ there a few common findings followed by many more relatively rare conditions. In chest radiography, diagnosis is both long-tailed and multi-label problem, patients often present with multiple simultaneously. While researchers have begun to study the problem of learning in recognition, studied interaction label imbalance co-occurrence posed long-tailed, disease...