Table 2.

Augmentations applied to the training set during the finetuning of Cellpose models.

Augmentation method	Parameter(s)	Parameter(s) distribution	Intent
Scaling	Scaling factor		Expand dataset with invariance to cell size
Rotation	Rotation angle		Expand dataset with invariance to cell orientation
Flipping	Probability and		Expand dataset with invariance to cell orientation
Additive White Gaussian Noise (AWGN)	Mean and standard deviation		Emulate variation in the expression of the lit-up fluorescent pixels and random background noise( 13 )
Poisson Noise	—	—	Emulate the noise generated by the fluorescence\microscope imaging( 14 , 15 )
Salt and pepper	, , and		Emulate variation in the location and expression of the FLs by simulating activation or inhibitions of fluorescent proteins( 16 )
Brightness	Shift		Emulate the variance in microscope image acquisition and fluorescence intensity( 17 )