michalfaber/keras_Realtime_Multi-Person_Pose_Estimation

Logic behind the numbering scheme in mapIdx

Closed this issue · 1 comments

Hi, thank you for the openpose implementation in Keras. I am trying to train on custom data and check the output of the model so trained. I need to ignore some of the body parts and limbs due to limitations of the custom dataset. To do that, I need to understand 'limbSeq' and 'mapIdx' structure. 'limbSeq' is fairly easy to understand but I am unable to figure out the reasoning behind the pair of numbers present in mapIdx. If someone could please suggest it to me or point me towards an explanation, I would be grateful. Thanks!

@Naykira hey ,have you know that?i also cant understand them.do you have some time to point me towards an explanation? thank you very much