Automatic Liver Segmentation on Volumetric CT Images Using Supervoxel-Based Graph Cuts

Table 3

Examples of reduction of the volume size: train = Sliver07-train datasets; test = Sliver07-test datasets.

Data	Original CT [voxels]	Resampled CT [voxels]	Abdominal region [voxels]	Liver VOI [voxels]	VOI supervoxels [supervoxels]
Train number 5	512 × 512 × 319	197 × 197 × 212	163 × 175 × 200	121 × 148 × 192	40 × 49 × 64
Train number 11	512 × 512 × 388	200 × 200 × 258	172 × 136 × 148	172 × 136 × 148	57 × 45 × 49
Train number 14	512 × 512 × 129	245 × 245 × 427	217 × 162 × 138	137 × 160 × 138	45 × 53 × 46
Test number 1	512 × 512 × 502	253 × 253 × 267	216 × 219 × 122	129 × 183 × 122	43 × 61 × 40
Test number 4	512 × 512 × 165	253 × 253 × 328	234 × 202 × 176	143 × 178 × 169	47 × 59 × 56