Deformable DETR with Box Refine

JSON →
SenseTime vision
image

Deformable attention detection transformer with iterative box refinement for improved object localization.

vision
releasedMar 2021