Approaching expert results using a hierarchical cerebellum parcellation protocol for multiple inexpert human raters

► We present a system for manual delineation of the cerebellum by inexpert raters. ► A hierarchical protocol enables efficient and accurate results from inexperts. ► A "rapid review" step detects large errors. ► Statistical fusion of several parcellations improves robustness and accuracy. ► The quality of the resulting labels approaches that of a human expert.