Research Interests

Research themes grounded in publications and current work

My work centers on medical imaging and computer vision, with a long-term focus on building AI systems that are clinically meaningful, robust, and interpretable. The themes below summarize the directions that are explicitly supported by my publications, CV, and current Ph.D. work.

Interest 01

Medical Image Analysis and Representation Learning

I am interested in learning image representations that preserve clinically meaningful structure and support downstream medical decision-making.

Representative work: anatomically consistent embedding for chest radiography, BMVC 2023.

View related Scholar results

Interest 02

Medical Image Segmentation and Vision Architectures

I work on segmentation models and architecture design for medical images, especially when structure and efficiency matter.

Representative work: hybrid ladder transformers for medical image segmentation, MIDL 2022.

View related Scholar results

Interest 03

Data-Efficient Learning and Transfer in Medical Imaging

Another recurring direction is learning under limited supervision, including zero-shot transfer, multimodal pretraining, and other data-efficient strategies for medical imaging.

Current context: UZH work on zero-shot learning and multimodal pretraining, plus present interest in data-efficient learning.

View related Scholar results

Interest 04

Medical XAI, Trustworthy Evaluation, and Clinical Utility

My current Ph.D. direction emphasizes explainability and trustworthy evaluation, with the goal of making AI systems useful in real clinical workflows rather than only on benchmarks.

Current context: Ph.D. topic stated as medical XAI and trustworthy AI for medicine.

View related Scholar results