이상철 연구실

AI-based analysis of medical images including polyp segmentation, ECG classification, and cardiovascular disease detection using deep learning techniques.

Frequency-based methods for detecting and localizing image forgeries using multi-task learning and attention mechanisms.

Techniques for improving model generalization across different domains and modalities, including domain adaptation for medical image segmentation.

Methods for few-shot learning and classification of time series data using meta-learning and contrastive learning approaches.

Morphological and deep learning approaches for robust segmentation of cells and nuclei in microscopy images.

Frequency-based and corruption-robust data augmentation techniques for improving image classification performance.