MSGFormer: A DeepLabv3+ Like Semantically Masked and Pixel Contrast Transformer for MouseHole Segmentation.
Peng YangChunmei LiChengwu FangShasha KongYunpeng JinKai LiHaiyang LiXiangjie HuangYaosheng HanPublished in: IEEE Access (2024)
Keyphrases
- pixel classification
- foreground background
- pixel level
- homogeneous regions
- grey level
- image segmentation
- spatial coherence
- level set
- pixel wise
- segmentation method
- segmentation algorithm
- gray level images
- fully automatic
- power system
- object segmentation
- region growing
- image pixels
- graph cut segmentation
- fuzzy logic
- motion segmentation
- shape prior
- color image segmentation
- intensity variations
- pixel values
- segmentation accuracy
- medical imaging
- object detection
- input image
- region boundaries
- robust segmentation
- labeling problems
- motion estimation
- multiscale