cpsiff / plant-segmentation Goto Github PK

other sklearn metrics include the dice coefficient (as F1 score), which we could lump in with precision and recall while we're at it.

Should be able to perform denoise in python to apply to any (noisy) mask, and output a binary mask as is currently done in denoiseMask.m

single platform would limit the headache of switching platforms when processing image

look into feasibility using multidim gmm w/ intensity/color as a classification for foreground/background as way of segmenting plant

per text received 1:03 am:

"What if we used some generic segmentation method like from skimage or something and then we just select the segmented areas with a certain amount of plant color in them"

• Carter