1 paper · 1 filter
Kang Liu, Zhuoqi Ma, Siyu Liang +5
Despite recent advances in medical vision-language pretraining, existing models still struggle to capture the diagnostic workflow: radiographs are typically treated as context-agno…