0
$\begingroup$

If I train a multi label classifier on full images and then I feed some patches of these images will it accurately generate the labels which comes in that patch?

For example if I train an image classifier on a singly image having a size of 300 including elephant, cheetah and monkey labels.

Now during prediction I make this image a patch by removing other two animals and just keep elephant, will my multi label classifier predict that?

Improve this question
$\endgroup$

1 Answer 1

Reset to default
0
$\begingroup$

Assuming that your ML model learned the correct underlying distribution to identify elephant, cheetah and monkeys individually and your new (test) patch image belongs to that distribution, the answer would be yes.

Improve this answer
$\endgroup$
2
  • $\begingroup$ Is there any tutorial to further read about this? $\endgroup$
    – Tensor
    Commented Dec 21, 2022 at 13:39
  • $\begingroup$ My answer was generic (I didn't describe a specific ML algorithm). Have a look at how neural network classifiers learn distributions. $\endgroup$
    – desert_ranger
    Commented Dec 21, 2022 at 13:56

You must log in to answer this question.

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