Screening Ziehl-Neelsen (ZN)-stained sections for acid-alcohol-fast bacilli (AAFB) is laborious, and sparse bacilli are easily missed. This article presents an automatic screening algorithm using digital image analysis designed to assist human diagnosis of tissue sections. The algorithm uses multiderivative source potentiators and suppressors feeding into interconnected product nodes that result in a probability value for each image (the likelihood that it contains AAFB) and a spatial probability map showing the position of any bacillus. For the study, 3,000 images from ZN-stained tissues were captured, 1,000 were used to train the algorithm, and 2,000 were used to test it. The algorithm successfully ranked AAFB-containing images as the highest in the data sets, despite only single bacilli being present in sparse images (occupying 0.0024% of the image) and despite tissue and staining artifacts. These results suggest that this automated screening assistance method has the potential to save time and money, which is especially important in resource-poor health services.