Congratulations on the Paper “LVM-Med: Learning large-scale self-supervised vision models for medical imaging via second-order graph matching” (Authors: Duy M. H. Nguyen, Hoang Nguyen, Nghiem T. Diep, Tan N. Pham, Tri Cao, Binh T. Nguyen, Paul Swoboda, Nhat Ho, Shadi Albarqouni, Pengtao Xie, Daniel Sonntag, Mathias Niepert) were published in NeurIPS 2023.
Students and graduate students participating in the research include:
(1) Nguyễn Hồ Minh Duy, a former student of the Faculty of Mathematics – Computer Science and a MSc student in Machine Learning at Max-Planck Institute, Germany. (Co-first author);
(2) Nguyễn Minh Hoàng, the 4th year student of the Faculty of Mathematics – Computer Science (Co-first author);
(3) Cao Thiên Trí, an alumni of Advanced Program in Computer Science of the Faculty of Information Technology;
(4) Phạm Ngọc Tân, a Master’s student in Artificial Intelligence;
(5) Diệp Tường Nghiêm, the 2nd year student of the Advanced Program in Computer Science of the Faculty of Information Technology.
The paper focuses on designing pre-trained models for medical images. It can be challenging to structure models for medical images due to domain shifts. For instance, CT, MRI, and X-ray images have various characteristics that make it demanding to create a unified model. However, this paper gave out a solution called LVM-Med, which is known as a contrastive learning algorithm based on high-order graph matching. The algorithm mentioned (LVM-Med) accounted for approximately 6 to 7 % accuracy than other methods and applicable hard-solving problems such as brain tumour classification or diabetic retinopathy grading.
- Link to the paper: https://arxiv.org/abs/2306.11925