AFLW2000-3D translation

[asw91666]

Hello. Thanks for sharing code !

According to your paper, the translation performance of the img2pose model was evaluated using the AFLW2000-3D dataset (referenced in Table 1). However, to my understanding, the AFLW2000-3D dataset only provides annotations for head rotation and not for head translation. Could you kindly share how the translation was measured in your study?

I am looking forward to understanding this aspect of your work more deeply. Any clarification or additional information you could provide would be greatly appreciated.