Adding register tokens to Vision Transformers eliminates high-norm background artifacts and raises state-of-the-art performance on dense visual prediction tasks.
arXiv preprint arXiv:2209.08575 , year=
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CV 2verdicts
UNVERDICTED 2roles
background 1polarities
background 1representative citing papers
H3D-MarNet suppresses metal artifacts in kVCT via wavelet preprocessing and transforms to MVCT using a dual-path CNN-transformer architecture with attention fusion, reporting 28.14 dB PSNR and 0.717 SSIM on affected slices.
citing papers explorer
-
Vision Transformers Need Registers
Adding register tokens to Vision Transformers eliminates high-norm background artifacts and raises state-of-the-art performance on dense visual prediction tasks.
-
H3D-MarNet: Wavelet-Guided Dual-Path Learning for Metal Artifact Suppression and CT Modality Transformation for Radiotherapy Workflows
H3D-MarNet suppresses metal artifacts in kVCT via wavelet preprocessing and transforms to MVCT using a dual-path CNN-transformer architecture with attention fusion, reporting 28.14 dB PSNR and 0.717 SSIM on affected slices.