Some problems about dataloader
Opened this issue · 3 comments
Thank you for you code. I am now modify your data loader to migrate the boundary clue into our project. Would you mind telling me the meaning of chin, leb,reb,bon.breath and so on? Some name is difficult for us to understand. Thx a lot
I tried to define all the boundary by this rule, 'face_boundary','left_eye_bow','right_eye_bow','nose_bridge','nose_under','left_upper_eye','left_lower_eye','right_upper_eye','right_lower_eye','up_mouth_up_part','up_mouth_low_part','low_mouth_up_part','low_mouth_low_part'
Are these right annotation for datasets
chin--face_boundary
leb--left_eye_bow
reb--right_eye_bow
bon--nose_bridge
breath--nose_under
...
I use the first word to simplify the representation.
chin--face_boundary
leb--left_eye_bow
reb--right_eye_bow
bon--nose_bridge
breath--nose_under
...
I use the first word to simplify the representation.
THX,another small question.
how can we get 'lo_eye_corner_index_x' , Since the original dataset annotation has only 96 cordinate, how can we get the right index number.