./demos/image_vis3d_6DRepNet.py fails to predict.

[Redhwan-A]

I tried to predict HPE using images that include one person, but it fails to predict in both your DirectMHP and 6DRepNet.
Could you create code to predict the head pose for images that include one person, please?

[hnuzhy]

Sometimes, the model does fail to give a head pose of one person when his/her head can not be detected successfully. Try to change the confidence threshold of the head detector. If it still not work, you may consider abandoning your challenging cases (an image with low resolution, motion blur, poor lighting or etc.)