Automatic Labeling of Vertebral Levels Using a Robust Template-Based Approach

Table 1

Acquisition parameters for the tested spine MRIs. Voxel size was 1×1×1 mm³.

Contrast	Image size (voxel)	Visible vertebrae (range)	Subject
T1	264∗684∗174(#)	C1–L3	Subject 1
T1	272∗732∗172(#)	C1–L2	Subject 2
T1	265∗633∗174(#)	C1–L3	Subject 3
T1	268∗757∗166(#)	C1–L3	Subject 4
T1	264∗690∗174(#)	C1–L3	Subject 5
T1	265∗679∗173(#)	C1–L3	Subject 6
T1	264∗384∗175	C1–T7	Subject 7
T1	264∗384∗175	C1–T6	Subject 8
T2	384∗384∗52	C1–T5	Subject 8
T2	384∗384∗52	C1–T3	Subject 9
T2	384∗692∗51(#)	C1–L3	Subject 5
T2	384∗384∗144	C1–T2	Subject 3
T2	384∗384∗52	C1–T3	Subject 7
T2	386∗678∗51(#)	C1–L1	Subject 6
T2	288∗512∗52(#)	C1–T6	Subject 10
T2	384∗384∗52	C1–T7	Subject 11

(#) Two fields of view were acquired sequentially and then images were stitched together using offline reconstruction tools from the console.