Output

[waqarsqureshi]

output of the inference is the prediction probabilities of foreground and background, the x and y and theta, and the Lane prior. is it possible to generate an image mask from the foreground and background probabilities?