Completely Occluded and Dense Object Instance Segmentation Using Box Prompt-Based Segmentation Foundation Models.
Zhen ZhouJunfeng FanYunkai MaSihan ZhaoFengshui JingMin TanPublished in: CoRR (2024)
Keyphrases
- image segmentation
- object segmentation
- segmentation algorithm
- global shape
- segmentation method
- level set
- fully automatic
- medical images
- image analysis
- multiple objects
- segmentation accuracy
- object tracking
- pixel level
- motion segmentation
- displacement field
- multiscale
- signed distance
- range scans
- three dimensional
- fully unsupervised
- optical flow estimation
- intensity images
- object model
- shape prior
- region growing
- computer vision
- energy function
- probabilistic model
- active contours
- model selection
- graph cuts
- d objects