Dear author,
How to re-implement the tracker with VGG-M features as presented in the paper, I just want to test the deep version on our machine and dataset, Thank you!
I try to use CN feature, the extracted sample, however, still remains 31-dimensions instead of 10-dimensions, and there seems to lack some important params related to VGGM feature.
Dear author, How to re-implement the tracker with VGG-M features as presented in the paper, I just want to test the deep version on our machine and dataset, Thank you! I try to use CN feature, the extracted sample, however, still remains 31-dimensions instead of 10-dimensions, and there seems to lack some important params related to VGGM feature.
Best wishes!