GMIM: Self-supervised pre-training for 3D medical image segmentation with adaptive and hierarchical masked image modeling.
Liangce QiZhengang JiangWeili ShiFeng QuGuanyuan FengPublished in: Comput. Biol. Medicine (2024)
Keyphrases
- single image
- image features
- input image
- multiscale
- image data
- image retrieval
- image segmentation
- image analysis
- template matching
- image content
- image representation
- lighting conditions
- classifier training
- pixel values
- test images
- image regions
- image processing
- high resolution
- computer vision
- image classification
- similarity measure
- multiresolution
- low level
- image matching
- digital images
- region of interest
- keypoints
- binary images
- supervised learning
- image pixels
- object detectors
- linear predictors
- segmentation algorithm