Previous Contents Next
Previous : Purpose of the hyperparameters  Next : Choice of the estimator 
How to choose the right hyperparameters?
We can sometimes choose nearly correct  and  in an empirical way,
to resore edges avoiding to obtain a too noisy image.
Original image
Corrupted image
=0.3 =5
best result
(empirical choice)
The following pictures illustrate the difficulty of a manual hyperparameter choice :
=2 =1
=0.02 =5
insufficient regularization
=0.3 =0.5
too small threshold
=0.3 =50
too high threshold
An automatic choice of the  and  hyperparameters is necessary!
Previous Next
André Jalobeanu - 24 Aug 1998