Adrian Bulat

Hi @camel007 , I checked a bit the structure of the graph you posted and it seems to be correct (of course except for the missing conv layers). At the...

@camel007 also, if you have any particular questions regarding the architecture, feel free to ask.

@natanielruiz I am really sorry about the delay, I totally forgot about this issue. The above figure was missing the Binary Activation layers. I will try to upload it tomorrow

Hi @RichardWenz , Yes, it is possible if needed. Right now, for simplicity I am just using the ground truth points to create a bounding box and then compute a...

While unfortunately I don't have the time to look thru your code line-by-line at the moment, my comments are as follow (will have a look at a later point): -...

@thanhhvnqb I have run the script available online and I got 300W-3D: 81.1% Menpo-3D: 72.5% I am now away but I can send you this particular predictions obtained using this...

Thanks for reporting this. The demo should be able to run even without this, however I will fix this by the end of this week since there are a few...

Hi @PyJulie, I am glad it worked and you liked it! I will add one tomorrow.

Hi, apologies for this, I will push a new version tomorrow, in the meantime it could be pip installed from git directly.

In principle it's possible, there a few systems around that use this landmark detector for this (deep face-like). However, using it directly its sub-optimal, as one will need to train...