GeoTranMesh: a geometry-guided multi-branch mesh transformer for 3d liver segmentation

Puntos clave

• The aim is to enhance the segmentation of 3D liver anatomy using advanced geometric and attention-based techniques.
• Utilized a geometry-guided approach for segmentation
• Implemented hybrid attention mechanisms
• Employed directional multi-branch fusion techniques
• Achieved precise segmentation of complex liver structures
• Enhanced morphological feature learning
• Showed potential for clinical application in liver surgery and augmented reality environments