Definition of live/dead organoid about intestinal_organoid_yolov8 HOT 3 CLOSED

Hello Ajinkya, dying/overgrown organoids appear very dark under brightfield microscopy. This is because all the cells that would normally be shed into the intestinal lumen are being shed inside the organoid. So the YOLOv8 model we trained is picking those features. Morphology wise they are not so different from differentiated organoids until late stages where dead organoids lose their structural integrity and "disintegrate".

from intestinal_organoid_yolov8.

Ah ok - thanks for clarifying!
And to be a bit more specific - below which values of intensity do you consider the organoid dead? Say if you scale the image be lie between 0-255, what would be the threshold value? And how would/have you validated that particular value?

Thanks and regards,
Ajinkya

from intestinal_organoid_yolov8.

Hi again, there is not a defined value below which the researchers that annotated the ground truth images considered the organoids dead, but rather general appearance (morphology and darkness). Whenever I find some time I can implement an intensity measurement using skimage.measure and plot the grey values to get an idea of these for the 3 different classes. So far it has limited functionality since the main interest of the users is getting basic morphology and counts of spheroids and differentiated organoids.

from intestinal_organoid_yolov8.