1
$\begingroup$

so a Faster R-CNN consists of 3 parts:
-the Base CNN
-the Region Proposal Network with RoIpooling
-and the classification CNN at the end

So I asked myself if all these three Networks are trained seperatly and get united at the end. Or if the whole constillation is trained together (so that the input flows through all three Networks and so does the backpropagation)?

If the sencond is the case, how does the RPN fits into that? Because the output of the RPN with RoI-pooling are multiple smaller regions which then get forwarded one after another to the last CNN for classification and so it's no more a continous united CNN, right?

I would be really happy to get some answers because I somehow can't find any else where, thanks :)

Improve this question
$\endgroup$

0

Reset to default

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.