Masked pre‐training and pseudo‐vessel enhancement‐based semi‐supervised cerebral vascular segmentation

Key Points

• The aim is to develop a framework for effective cerebrovascular segmentation that requires minimal labeled data.
• Developed a semi-supervised framework for cerebrovascular segmentation
• Implemented masked pre-training techniques
• Enhanced small vessel continuity while reducing over-segmentation
• Evaluated robustness and generalizability in various contexts
• Achieved accurate segmentation of cerebrovascular structures
• Enhanced continuity of small vessels
• Reduced over-segmentation due to noise
• Demonstrated strong performance in real-world clinical scenarios