• Laser & Optoelectronics Progress
  • Vol. 57, Issue 18, 181024 (2020)
Wei Liu* and Hongwei Ge
Author Affiliations
  • Key Laboratory of Advanced Process Control for Light Industry, Ministry of Education,
  • show less
    DOI: 10.3788/LOP57.181024 Cite this Article Set citation alerts
    Wei Liu, Hongwei Ge. Fast Face Recognition Method Based on Sparse Representation[J]. Laser & Optoelectronics Progress, 2020, 57(18): 181024 Copy Citation Text show less
    Number of samples456789
    DSRC0.87340.90410.92330.94380.94590.9542
    Our algorithm +DSRC0.87380.90470.92340.94300.94650.9539
    CRC0.83000.86990.88520.90600.91350.9216
    Our algorithm +CRC0.85320.89030.90310.92200.92820.9363
    CFFR0.87090.89780.91530.91980.92740.9348
    Our algorithm +CFFR0.87650.90220.92280.92900.93620.9372
    Table 1. 0 Recognition accuracy of different algorithms on FEI database
    Number of samples23456
    Time by DSRC /s0.16360.18370.18930.18420.1673
    Time by our algorithm +DSRC /s0.14270.15800.16090.15310.1397
    Improvement /%12.814.015.016.916.5
    Time by CRC /s0.17530.18870.19530.18690.1681
    Time by our algorithm +CRC /s0.14850.16220.16330.15740.1440
    Improvement /%15.314.016.415.814.3
    Time by CFFR /s0.79120.93821.07801.23361.3746
    Time by our algorithm +CFFR /s0.60310.77520.82430.83040.9582
    Improvement /%23.717.423.532.730.3
    Table 1. Total CPU time used by different algorithms on ORL database
    Number of samples23456
    DSRC0.85860.91860.94690.96180.9699
    Our algorithm +DSRC0.85930.92100.94740.96370.9731
    CRC0.83310.89370.92480.94730.9569
    Our algorithm +CRC0.84290.89900.93080.94900.9597
    CFFR0.85250.91190.94120.95880.9675
    Our algorithm +CFFR0.85430.91490.94530.95930.9678
    Table 2. Recognition accuracy of different algorithms on ORL database
    Number of samples3456789
    Time by DSRC /s0.24410.25820.27220.27840.28380.27640.2562
    Time by our algorithm +DSRC /s0.20970.21980.22400.23300.24100.23310.2196
    Improvement /%14.114.917.716.315.115.714.3
    Time by CRC /s0.25740.27260.27740.29090.29010.28690.2771
    Time by our algorithm +CRC /s0.20530.22230.22940.24950.24790.24590.2377
    Improvement /%20.218.517.314.214.514.314.2
    Time by CFFR /s1.14951.32331.61041.89531.90212.16212.2487
    Time by our algorithm +CFFR /s0.79861.07431.23991.34401.49181.51371.4171
    Improvement /%30.118.823.029.121.630.037.0
    Table 3. Total CPU time used by different algorithms on GT database
    Number of samples3456789
    DSRC0.60160.65640.69200.72290.75320.77350.7938
    Our algorithm +DSRC0.60250.65870.69400.72460.75450.77650.7944
    CRC0.55600.60520.64000.66770.68630.70910.7255
    Our algorithm +CRC0.57540.62750.66300.69140.70790.73160.7461
    CFFR0.60050.64220.67980.69600.71850.73510.7567
    Our algorithm +CFFR0.60260.65350.69190.71670.73550.75170.7695
    Table 4. Recognition accuracy of different algorithms on GT database
    Number of samples2345
    Time by DSRC /s2.41902.38532.03321.5536
    Time by our algorithm +DSRC /s1.85431.97831.59011.2123
    Improvement /%23.317.121.821.0
    Time by CRC /s2.54012.44872.08051.6036
    Time by our algorithm +CRC /s1.93781.85011.62141.2563
    Improvement /%23.724.422.121.7
    Time by CFFR /s20.80629.56330.30925.937
    Time by our algorithm +CFFR /s11.69316.08017.34915.313
    Improvement /%43.845.642.841.0
    Table 5. Total CPU time used by different algorithms on FERET database
    Number of samples2345
    DSRC0.65800.61630.79000.8050
    Our algorithm +DSRC0.66160.61790.78800.8075
    CRC0.58500.50880.56330.7050
    Our algorithm +CRC0.59410.51940.58350.7218
    CFFR0.62500.54000.64000.7400
    Our algorithm +CFFR0.61870.53800.65900.7408
    Table 6. Recognition accuracy of different algorithms on FERET database
    Number of samples6810121416
    Time by DSRC /s7.44617.97009.80299.21908.73807.9250
    Time by our algorithm +DSRC /s4.87715.34896.42906.25806.25205.8150
    Improvement /%34.532.934.432.128.526.6
    Time by CRC /s7.61648.04599.97909.76419.04468.1810
    Time by our algorithm +CRC /s5.00305.37946.50996.62226.52536.0345
    Improvement /%34.333.134.732.227.826.2
    Time by CFFR /s99.132121.367135.631148.894153.322150.028
    Time by our algorithm +CFFR /s46.31457.44363.90471.82381.38082.083
    Improvement /%53.252.752.951.846.945.3
    Table 7. Total CPU time used by different algorithms on AR database
    Number of samples6810121416
    DSRC0.66170.63750.59220.73630.89030.9192
    Our algorithm +DSRC0.67210.64510.60140.73780.89510.9200
    CRC0.64960.64720.60830.63450.82640.8692
    Our algorithm +CRC0.66000.65120.61310.66760.85480.8808
    CFFR0.70710.70420.66250.73990.89440.9167
    Our algorithm +CFFR0.69080.69630.65890.75540.90900.9267
    Table 8. Recognition accuracy of different algorithms on AR database
    Number of samples456789
    Time by DSRC /s0.82430.85121.08871.08850.97730.8304
    Time by our algorithm +DSRC /s0.44510.46410.59730.63690.57390.5003
    Improvement /%46.045.545.141.541.339.8
    Time by CRC /s0.83510.83351.07131.02010.92030.8090
    Time by our algorithm +CRC /s0.48620.47060.58710.59300.55580.4892
    Improvement /%41.843.545.241.939.639.5
    Time by CFFR /s4.94696.24307.22777.57827.39077.1898
    Time by our algorithm +CFFR /s2.21112.66023.01813.05233.14422.8095
    Improvement /%55.357.458.259.757.560.9
    Table 9. Total CPU time used by different algorithms on FEI database
    Wei Liu, Hongwei Ge. Fast Face Recognition Method Based on Sparse Representation[J]. Laser & Optoelectronics Progress, 2020, 57(18): 181024
    Download Citation