FedDM: Federated Weakly Supervised Segmentation via Annotation Calibration and Gradient De-Conflicting.
Meilu ZhuZhen ChenYixuan YuanPublished in: IEEE Trans. Medical Imaging (2023)
Keyphrases
- weakly supervised
- semantic segmentation
- weakly labeled
- superpixels
- fully supervised
- segmentation algorithm
- object class
- relation extraction
- image segmentation
- object segmentation
- topic models
- level set
- multiscale
- segmentation method
- shape prior
- semi supervised
- energy function
- active learning
- object detectors
- object recognition
- metadata
- object detection
- pairwise
- video sequences