aand ? thresholded.shape
is what ?
however, before we try to fix your current errors – what exactly did you train your network on ? color images ? then you must use the same for prediction, so most probably you should remove all that preprocessing
aand ? thresholded.shape
is what ?
however, before we try to fix your current errors – what exactly did you train your network on ? color images ? then you must use the same for prediction, so most probably you should remove all that preprocessing