0
$\begingroup$

I would like to classify whether a pot of water is boiling or not using a CNN.

Is it enough to take pictures of boiling water using only one pot, or should I use different pots for this to generalize well?

$\endgroup$

1 Answer 1

0
$\begingroup$

It depends of your application. If your aim is to have a general boiling detector, yes, if not, don't bother yourself.

I would suggest, if you can, to crop the picture of where the water is, to not be inflenced by the overall context.

I am pretty sure you will still have good results even with one pot. But you may see some accuracy drop when using different color pot during test time (if that color haven't been seem during training). Some data augmentation with color tweaking will help. That will redirect the network into texture analysing rather than color analysing.

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .