MeDSLIP: medical dual-stream language-image pre-training with pathology-anatomy semantic alignment

Mar 15, 2024·
Wenrui Fan
Wenrui Fan
,
Mohammod Naimul Islam Suvon
,
Shuo Zhou
,
Xianyuan Liu
,
Samer Alabed
,
Venet Osmani
,
Andrew J. Swift
,
Chen Chen
,
Haiping Lu
· 0 min read
MeDSLIP overview
Abstract
MeDSLIP is a dual-stream vision-language pre-training framework that models anatomy- and pathology-aware fine-grained alignment in medical images and reports.
Type
Publication
arXiv
Status
Open access
publications
Vision-Language Learning Medical AI Multimodal Learning
Wenrui Fan
Authors
Wenrui Fan (he/him)
AI Research Engineer & PhD Student
Wenrui Fan is an AI Research Engineer and PhD student at the University of Sheffield, supervised by Prof Haiping Lu. His research focuses on multimodal AI, AI for science, agentic AI, AI for healthcare, biomedical AI, and AI-driven protein design.