Not All Images are Worth 16x16 Words: Dynamic Vision Transformers with Adaptive Sequence Length.
Yulin WangRui HuangShiji SongZeyi HuangGao HuangPublished in: CoRR (2021)
Keyphrases
- image data
- image database
- image retrieval
- image features
- image analysis
- input image
- image understanding
- ground truth
- image classification
- three dimensional
- image collections
- multiple images
- image regions
- rigid body
- edge detection
- image registration
- image processing
- computer vision
- image set
- fixed length
- segmentation algorithm
- dynamic environments
- feature points
- segmentation method
- computer graphics
- test images
- image annotation
- super resolution
- vision system
- color images