Figure 8 in your paper

Hi, I see you have resized all FPHAB images to [480x270] dimension in [this code ](https://github.com/hassony2/handobjectconsist/blob/master/reduce_fphab.py)

In figure 8, are the reported pixel erros in the size of [480x270], or the full resolution size? I'm working on a research project and want to compare with this method fairly.