Rethinking and Improving Relative Position Encoding for Vision Transformer.

Kan Wu Houwen Peng Minghao Chen Jianlong Fu Hongyang Chao

Published in: CoRR (2021)

Keyphrases

relative position
geometric properties
computer vision
image processing
real time
fuzzy logic
vision system
spatial relationships
image segmentation
three dimensional
optical flow
similarity search
power system
fault diagnosis