|
| Study | MR sample | Number of subjects | Methods used | Efficiency measures |
|
| [47] | 3D DESS | 33 subjects | Region growing and seed selection | DSC 82.8%, 83.1%, and 72.6% for femur, tibia, and patella, respectively |
| [15] | T1-weighted 3D DESS | 20 subjects | Side selection and coupled cartilage segmentation | DSC 0.80
Sensitivity 90.0
Specificity 99.8 |
| [48] | 3D DESS | 20 subjects | Spatial fuzzy C-mean | DSC of femur, tibia, and patella are 87.1, 81.1, and 84.8, respectively |
| [37] | DESS | 50 subjects | Statistical shape model | Accuracy 74 ± 5
DSC 86.1% |
| [49] | 2D scan | 100 subjects | Random walker algorithm | DSC 0.8758 |
| [50] | DESS | 8 subjects | DRLSE algorithm | DSC 90.28
Sensitivity 91.14
Specificity 99.12 |
| [35] | 3D DESS | 10 subjects | K-nearest neighbor | Sensitivity 84.2
Specificity 99.9
Dice volume 0.81 |
| [34] | 3D image | 139 subjects | Voxel classification | DSC 0.80 ± 0.04 |
| [51] | DESS | 13 subjects | Hierarchical classifier and random forest | Surface positioning errors femur 0.03 ± 0.19 tibia 0.10 ± 0.17 |
| [52] | T1-weighted 3D SPGR and T2-weighted GRE | 9 subjects | Three-label segmentation | Median dice coefficient femur 0.935 and tibia 0.938 |
| [53] | 3D SPGR | 155 subjects | Longitudinal three label segmentation | Femur DSC 75%
Sensitivity 78%
Specificity 99.9% |
| [54] | Sagittal DESS | 88 subjects | Convolution neural network | DSC 83.8% |
| [55] | DESS | 40 subjects | Active appearance model (statistical model) | RMSE for femur 1.49, tibia 1.21 |
| [56] | PDW TSE | 14 subjects | Distance-weighted directional gradient | Intensity level, 2∼4 |
| [57] | T1-weighted | 30 subjects | Multiatlas segmentation constrained graph | Femur average surface distance 0.36 mm |
| [58] | 2D ASM | 140 subjects | Rigid multiatlas registration | Dice volume tibia 0.8 and femur 0.87 |
| [59] | T1-weighted SPGR | 20 subjects | Statistical method | DSC of femur 0.871, tibia 0.852, and patella 0.645. Sensitivity of femur 0.947, tibia 0.949, and patella 0.909. Specificity of femur 0.988, tibia 0.993, and patella 0.996. |
| [60] | T1-weighted FS | 14 subjects | Validation of statistical method | Meniscus DSC 0.75, sensitivity 0.72, and specificity 1.00 |
| [61] | T1-weighted | 100 subjects | Coarse to fine method with min-cut | DSC of femur 0.947, and tibia 0.968 |
|
