This research introduces a new automated method for precise Couinaud liver segmentation using contrast-enhanced MRI images by identifying seven anatomical landmarks, improving surgical planning and reducing complications.
By implementing a multi-task learning framework, the study syncs landmark detection with segmentation, achieving a high average Dice Similarity Coefficient (DSC) of 85.29%, outperforming previous models.
The clinical application of this technique may lead to more personalized surgical plans, decreased operative risks, and better overall patient outcomes by preserving healthy liver tissue.