Luca Bertinetto

Hi @LinHungShi, Sorry for the very late answer but I took a break from my PhD to do an internship at the moment. 1) The concept of **scale** is related...

Hi, Thanks for the interest in our work. We are planning to release the training code towards the end of May.

I think you are using a version of matconvnet that is too old. Try to use something >beta21 and check that dagnn.Conv implementation of the version you pick has the...

Just define arrays of suitable values and take one random index per array per iteration.

Hi, I can't really tell. Just make sure the images are actually there :)

Hi, This is the protocol we adopted, which we used to avoid overfitting to the test set(s). * We train CFNet/SiamFC for 100 epochs * We measure its *tracking* performance...

Hi, Can you detail your problem a bit more?

With the XCorr layer, it is convenient to search over candidates of same size of the target. Regressing to width and height would require a different architecture.

I think we have some problems with newer versions of matconvnet/matlab/cuda/cudnn :\ Soon I will try to get the exact versions we use and report them on README, but I...

Hi! We always wanted to try, but we never actually did it. It would be interesting for example to see how it works for pedestrians or faces, both classes that...