Deep Learning
Deep Learning & Neural Networks
Domain Focus & Clinical Overview
Advanced deep network architectures, vision transformers, self-supervised representation learning, and model optimization for medical intelligence.
Core Methodologies & Technical Scope
Core Research Pillars:
- 3D Masked Autoencoder Pre-training on volumetric CT/MRI.
- Lightweight Convolutional-Transformer Hybrid Architectures.
- Quantization & TensorRT Inference Acceleration for Hospital Edge Servers.
- Self-supervised contrastive representation alignment.