Adapting Pre-trained Vision Transformers from 2D to 3D through Weight Inflation Improves Medical Image Segmentation.
Yuhui ZhangShih-Cheng HuangZhengping ZhouMatthew P. LungrenSerena YeungPublished in: ML4H@NeurIPS (2022)
Keyphrases
- medical image segmentation
- pre trained
- medical images
- training data
- curve evolution
- computer vision
- level set method
- medical image analysis
- medical imaging
- image segmentation
- statistical shape model
- training examples
- deformable models
- level set
- data sets
- multi view
- active contour model
- image analysis
- viewpoint
- three dimensional
- image processing
- appearance variations
- control signals
- neural network