关键词：荧光图像；肺结核；诊断；医疗器械
摘 要：Our algorithm applies morphological operations and template matching with a Gaussian kernelto identify TB-object candidates.Our object-evel classi cation is highly accurate, with Average Precision of 89:2%  2:1%. For slide-levelclassi cation, our algorithm performs at the level of human readers, demonstrating the potential for making a signi cant impact on global healthcare.