Table 1 Journal papers on classification and regression of finger movements using electromyography
From: Proportional estimation of finger movements from high-density surface electromyography
Ref. | Year | Classifier | Features | Finger moves | Subjects | Window (ms) | Electrodes | Accuracy | Â |
|---|---|---|---|---|---|---|---|---|---|
[38] | 2002 | kNN | DFT, AR | F (T,I,M-R-L) | ND (4) | - | 3 | 98Â % | Â |
[39] | 2009 | ANN | TD | F-E (T,I,M,R,L) | TR (1) ND (5) | 200 | 32 | 90Â % | Classification |
[18] | 2010 | kNN | MAV | F (T,I,M-R-L) | TR (1) | 250 | 16 | 86Â % | |
[68] | 2010 | EPM | TD | F (T,I,M,L,R) | ND (2) | - | 4 | >97Â % | |
[69] | 2011 | kNN | MAV | F (T,I,M,L,R) | TR (5) ND (5) | 250 | 8 | 79Â % (TR) 89Â % (ND) | |
[70] | 2012 | SVM, kNN | TD, AR | F (T,I,M,L,R) | ND (8) | 250 | 2 | 90Â % | |
[14] | 2012 | LDA, SVM, GMM | TD, AR | F-E (T,I,M-R-L) | PS (12) | 256 | 89 | 95Â % | |
[15] | 2013 | LDA, SVM | TD, AR | F-E (T,I,M-R-L) | ND (10) TR (6) | 200Â ms | 12 11 | 98Â % (ND) 90Â % (TR) | |
[41] | 2014 | KRLS | TD | F (T,I,M,R,L) | ND(40) | 100–400 | 12 | 90 % | |
[42] | 2015 | LDA | TD, AR | F (T,I,M) | ND(7) | 250 | 5 (iEMG) | 85Â % | |
[71] | 2006 | ANN | ENV | F (T,I,M,L,R) | TR(2) | - | 8 | (JA) Norm RMS error 8–20 % | Regression |
[17] | 2009 | ANN | RMS | F-E (I) | ND (15) | 100 | 1 | (JA) RMS error 0.085 rad −0.163 rad | |
[19] | 2012 | ANN | WL | F-E (T,I,M-R-L) | ND (5) | 32 | 4 | Norm RMS error 7–14 % | |
[20] | 2014 | ANN, GP | EMD | F-E (T,I,M-R-L) | ND (10) | - | 8 | (JA) Mean CORR 0.85 ± 0.07 (MCP) 0.78 ± 0.06 (PIP) 0.73 ± 0.04 (DIP) | |
[37] | 2014 | ANN | ENV | F-E (I,M,L,R) | ND (8) | - | 14 – 16 | (JA) R2 = 0.8 | |
[22] | 2014 | RR | RMS | F (T,I,M,R,L) | ND (10) | 200 | 10 | (FF) Norm RMS error 16Â % | |
[23] | 2014 | RR | ENV | F-E (I,M-R-L) | ND (10) | - | 10 | (FF) Norm RMS error 10–20 % |