0
$\begingroup$

I am trying to train a classifier (let's say to classify an object X or not X). But I don't have too much real images of object X which I want to classify. So I made some synthetic images of my own using some photo editing tools, but the problem is after feeding synthetic data my classifier doesn't recognise the real object. And since I had like 2-3 images of real object so I can't feed it more real data. Is there any solution on how to approach such problem with synthetic data to recognise real ones?

$\endgroup$
0
$\begingroup$

Neural networks require a lot of data for training. This is one of their largest drawbacks, and is very hard to get around this.

Using artificial data is not a solution. We don't really understand what the network is learning and we can't tell if it learned anything that is particular to the synthetic images.

What you can do is to use some sort of data augmentation. Rotate, scale, translate your images many times, and create a larger dataset.

Since it's such a small dataset, and if more data is really not possible to obtain, you should consider other approaches to help the classification. Either manually design some features to help a bit the classifier or do the classification with more traditional methods.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.