Zhi-Chao Tong

Diagnostic Performance of a Locally Deployed Vision Language Model for Bone Tumor Diagnosis Using Smartphone-Captured Images: Exploratory Retrospective Study

Adding the evaluated RAG configuration did not improve diagnostic accuracy and was associated with text-related diagnostic errors under smartphone-captured degraded imaging conditions, suggesting that persona design may influence the robustness of VLM responses to retrieved information, but this observation requires va...