Automatic segmentation of the pulmonary lobes with a 3D u-net and optimized loss function
This work addresses lung lobe segmentation for medical imaging, but it is incremental as it builds on existing u-net methods with an optimized loss function.
The researchers tackled the challenge of fully-automatic lung lobe segmentation by training a 3D u-net with a weighted Dice loss function, achieving a mean distance of 1.46 mm compared to 2.08 mm without weighting.
Fully-automatic lung lobe segmentation is challenging due to anatomical variations, pathologies, and incomplete fissures. We trained a 3D u-net for pulmonary lobe segmentation on 49 mainly publically available datasets and introduced a weighted Dice loss function to emphasize the lobar boundaries. To validate the performance of the proposed method we compared the results to two other methods. The new loss function improved the mean distance to 1.46 mm (compared to 2.08 mm for simple loss function without weighting).