You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi,
I tried to extract the features by extract_feature_rgb_full.py that you provided script in issue. I noticed that there was "OBJECT_" + side+"_FEAT" key in data.lmdb, but I checked 100DOH, which is used for detecting hands and object in the paper, but they seemed not to discriminate object sides (i.e., left/right objects). Could you give some instructions about how you divide the object bounding boxes as "left" and "right"? Really appreciate your help!
The text was updated successfully, but these errors were encountered:
Hi, thanks for your interest in our work. For the side of objects, it refers the object the corresponding hand is touching. For example, there could be left hand touch the object, but right hand not touch anything, or both hand touch the same object, ...
You could compute it using the IoU of hand-object bbox to determine the interaction object touched by left/right/both hands.
Hi,
I tried to extract the features by extract_feature_rgb_full.py that you provided script in issue. I noticed that there was "OBJECT_" + side+"_FEAT" key in data.lmdb, but I checked 100DOH, which is used for detecting hands and object in the paper, but they seemed not to discriminate object sides (i.e., left/right objects). Could you give some instructions about how you divide the object bounding boxes as "left" and "right"? Really appreciate your help!
The text was updated successfully, but these errors were encountered: