Figure 1.

Workflow of the automated atlas‐segmentation approach. A dedicated server monitors files acquired on the CT scanner and starts the segmentation algorithm once a new dataset is detected. Once completed, results are inserted and attached to the CT scan in the clinical database through the DICOM protocol. The blue insert details steps of the proposed segmentation procedure. Multiple atlases are initially matched to the patient dataset using a deformable registration technique. The STAPLE algorithm is then employed to compact individual atlas result into a probability map of organ position and shape. This estimate is further refined using level set‐based segmentation.