Revision history [back]

all the BackgroundSubtractor classes need to build a model over time, and thus need many images to learn from.

if you only have a single image, the keyword is "segmentation", you could e.g. try watershed or grabCut