I think the dataset you use to do experiments is not very convincing.
The 300W-LP dataset and AFLW2000 dataset are labeled by 3DDFA method(for head pose), and in your experiment you compare your results with 3DDFA when evaluating your method in AFLW2000 dataset. That is confusing!
I think the dataset you use to do experiments is not very convincing. The 300W-LP dataset and AFLW2000 dataset are labeled by 3DDFA method(for head pose), and in your experiment you compare your results with 3DDFA when evaluating your method in AFLW2000 dataset. That is confusing!
Can you help me answer my question?