Table 1.
Knee joint biomechanical gait data classification-related studies.
| Study | Pathology | Population | Biomechanical variables | Data acquisition | Feature Ext/select | Classification | Acc |
|---|---|---|---|---|---|---|---|
| [64] | ACL | 20 ACL 26 injured knees |
EMG | PDP 11/23 minicomputer | Fourier transform | K-means | — |
| [51] | OA | 50 OA 63 AS |
Spatiotemporal data Kinematics Kinetics |
Optoelectronic measurement system Force plate |
Global rep. PCA |
LDA | 94% |
| [38] | OA | 35 OA 107 AS |
Kinematics | 3D knee analyzer | Local rep. | ANOVA | — |
| [39] | OA | 12 OA 7 AS |
Acceleration data Kinematics Kinetics |
Vicon motion analysis system Force sensor integrated treadmill triaxial gyroscopes and accelerometers |
Local rep. | LDA | 87.9% |
| [40] | OA | 15 OA 15 AS |
Spatiotemporal data Kinematics Kinetics |
Optoelectronic measurement system Force plate |
Local rep. | DST | 96.7% |
| [41] | OA | 20 OA 25 AS |
Spatiotemporal data Kinematics Kinetics |
Optoelectronic measurement system Force plate |
Local rep. | DST | — |
| [74] | OA | 110OA 91 AS |
Spatiotemporal data Kinematics Kinetics |
Vicon motion analysis system | — | MLP | 98.5% |
| [52] | OA | 50OA 63AS |
Kinematics Kinetics |
Optoelectronic measurement system Force plate |
Global rep. PCA |
LDA | 92% |
| [42] | PFPS-OA | 13PFPS 5OA 20 PKR |
Kinetics | Kistler force platform | Local rep. Hill-climbing algorithm |
SVM | 85-92% |
| [69] | OA | 11 OA 12 AS |
Spatiotemporal data | Vicon motion analysis system GAITRite system |
Hill-climbing algorithm | SVM | 94.2% |
| [53] | OA | 20 OA 22 AS |
Kinematics Kinetics |
Optoelectronic measurement system Force plate |
Global rep. PCA |
DST | 97.62% |
| [60] | OA | 26OA 16AS |
Kinetics | Kistler force platforms | Global representation WT |
NN | 91% |
| [59] | OA | 26 OA 16 AS |
Kinetics | Kistler force platforms integrated to an ADAL treadmill | Global rep. WT |
NNC | 90% |
| [68] | OA | 11 OA 12 AS |
Spatiotemporal data | Vicon motion analysis system GAITRite system |
Hill-climbing algorithm | SVM | 88.89% |
| [70] | OA | 24 OA 12 AS |
Kinetics | Bertec platform | Wavelet packet based on the Fuz-Coc | DT-SVM | 93.44% |
| [37] | OA | 30 OA 14 AS |
Kinematics | KneeKG | Local rep. Global rep. |
SVD | 77.27% 93.18% |
| [61] | ACL | 29 ACL 15 AS |
Kinematics Kinetics |
ADAL treadmill Vicon motion analysis system |
Global rep. WT |
NNC | 83.2% |
| [49] | OA | 30 OA 14 AS |
Kinematics | KneeKG | Global rep. | SVD | 93.1% |
| [56] | ACL-R | 6 ACL-R 10 AS |
Kinematics | Four cameras motion analysis system (Innovision) | Global rep. PCA |
LR | 93.75% |
| [43] | OA | 18 KS 20 non-KS |
Kinematics | KneeKG | Local rep. | Student's t-test | — |
| [57] | AS | 111 AS | Kinematics | KneeKG | Global rep. PCA |
Discriminant model based on PCs' sign | — |
| [66] | OA | 2,900 OA | Spatiotemporal data | Computerized walking mat | — | CART | 89.5%–90.8% |
| [67] | OA | 2,911 OA | Spatiotemporal data | Computerized walking mat | — | CART | 89.5%–90.8% |
| [72] | OA | 47 OA 133 AS |
Kinetics | Kistler force plates | Probabilistic PCA | Bayes classifier | 82.62% |
| [73] | OA | 25 KS 25 non-KS 29 OA |
Kinematics | KneeKG | — | Bayes classifier | — |
| [46] | OA | 44 S 40 non-S |
Kinematics | KneeKG | Local rep. | CART | 84.7% |
| [48] | OA | 100 OA | Kinematics | KneeKG | Local rep. | RTs | 88% |
| [71] | ACL | 7 ACL 7 AS |
Kinematics | Vicon motion analysis system | PCA | SVM | 100% |
| [47] | OA | 100 OA 40 AS |
Kinematics | KneeKG | Local rep. | RTs | 85% |
| [65] | AS | 165 AS | Kinematics | KneeKG | Isometric mapping | DBSCAN algorithm | — |
| [62] | OA | 63 OA | Kinematics | KneeKG | WT | Kohonen neural network | 90.47% |