jspta - real-time facial recognition system-design implementation and validation

Upload: sathiskumaar

Post on 02-Jun-2018

222 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/10/2019 JSPTA - Real-Time Facial Recognition System-Design Implementation and Validation

    1/18

    J S P T A

    (2013) 1: 118:10.7726/.2013.1001

    R A

    ______________________________________________________________________________________________________________________________*C : _@.

    P, D. I T, D. A I T, B

    560056, I1

    ,

    M. M*

    R 29 A 2012; P N 10, 2012

    T () 2012. P .

    T , D S P (DSP)

    P . T , , . T

    WC, P C A

    (PCA) D C T (DCT). I , PC MATLAB. N, TMS320C6713DSP . S

    , . F,

    PC . A, DCT PCA .

    PCA; DCT; R S; DSP

    1.

    I ,

    . F

    , , , ,

    , , (Z , 2003; P , 1997).

    T . A ,

    ,

    ,

  • 8/10/2019 JSPTA - Real-Time Facial Recognition System-Design Implementation and Validation

    2/18

  • 8/10/2019 JSPTA - Real-Time Facial Recognition System-Design Implementation and Validation

    3/18

    Dr. M. Meenakshi / Journal of Signal Processing Theory and Applications (2013) 1: 1-18

    3

    S 2 / ,

    , S 3

    PCA DCT . T

    3. N, S 4

    . T TMS 320 C 6713 DSP S 5.

    T DSP 5. F

    6.

    .

    I .I , ,

    , . G , , ,

    / .

    T (C ., 1995).:

    1. F 2. H

    : F

    . S

    , , . T

    .

    F , ,

    . H, ,

    , .

    F

    . S

    , , ,

    (K, 1970). T

    , , S (S ., 1969). T

    . A ,

    K (1970).

    M , G . (1990)

    , Y . (1989). T

  • 8/10/2019 JSPTA - Real-Time Facial Recognition System-Design Implementation and Validation

    4/18

  • 8/10/2019 JSPTA - Real-Time Facial Recognition System-Design Implementation and Validation

    5/18

    Dr. M. Meenakshi / Journal of Signal Processing Theory and Applications (2013) 1: 1-18

    5

    .

    F 1 ,

    . I , ;

    . T . S

    . T

    ATM .

    Digital signal processingKit (DSK 6713)

    DOORDOORDOORDOOR

    LOCK

    OPENS ON

    IMAGE

    MATCH

    WEBCAM

    IM

    A

    G

    E

    1. F

    O

    .B

    . C

    , , , ,

    . R F R M

    , ,

    , , .

    A , . T PCA, ICA, L D A (LDA), E P (EP),

    EBGM, K M, T T, A A M (AAM), 3D M M,

    3D F R B F, SVM, HMM. T PCA DCT . S

    .

    ()

    PCA

    . T . D , , PCA. T

  • 8/10/2019 JSPTA - Real-Time Facial Recognition System-Design Implementation and Validation

    6/18

    Dr. M. Meenakshi / Journal of Signal Processing Theory and Applications (2013) 1: 1-18

    6

    PCA ,

    E. E

    . F

    . O

    F. 2:

    .PCA

    F PCA , T P

    T P, F. 2. PCA F. 3.

    T T P :

    S 1:

    1.

    T E

    2.

    C W

    3. R X, X WX

    S 2:

    C WX X W . F :

    1. C D W WX

    .2.

    C D . I D , ,

    WX .U , X . O, X ,

    WX . N .

  • 8/10/2019 JSPTA - Real-Time Facial Recognition System-Design Implementation and Validation

    7/18

    Dr. M. Meenakshi / Journal of Signal Processing Theory and Applications (2013) 1: 1-18

    7

    .F PCA

    Original faces Training set

    Start

    E = Eigenfaces (Training Set)

    W = Weights (E, Training Set)

    Input Unknown Image X

    WX= Weights (E, X)

    D = ave(distance(W, WX))

    D

  • 8/10/2019 JSPTA - Real-Time Facial Recognition System-Design Implementation and Validation

    8/18

    Dr. M. Meenakshi / Journal of Signal Processing Theory and Applications (2013) 1: 1-18

    8

    T PCA. T :

    S1: E : I ,

    () .L ( =1, 2,...,) (N2).

    S2: C .H, , :

    =

    =

    n

    i

    in 1

    1

    .

    S3: C . C [1, 2, 3, .. ]

    , =

    S4: C (C) = AAT.I , (V )

    . T , ..

    . C :

    nC

    T

    nn

    TT +++=

    ...................2211

    N C N2 N2, N2E

    . A ,

    . I N2 , , .., AAT, :

    ( )( ) ( )

    ( ) ( )iiiT

    iii

    T

    iii

    T

    AVAVAA

    VAVAAA

    VVAA

    =

    =

    =

    () ,

    . T, . S :

  • 8/10/2019 JSPTA - Real-Time Facial Recognition System-Design Implementation and Validation

    9/18

    Dr. M. Meenakshi / Journal of Signal Processing Theory and Applications (2013) 1: 1-18

    9

    [ ] =

    =

    ==

    n

    k

    kik

    i

    n

    i

    k

    i

    nii v

    v

    v

    v

    AVU1

    1

    2,1 .........

    S 4: R .

    F N , N0 , . T

    , .

    E ,

    . D N0 , V ,

    . F .

    ( )nk

    UW

    i

    ii

    T

    k

    i .,.........2,1; =

    =

    . N ( , )

    , (2, ) , .

    T () ( ). T

    .F, . T

    ,T

    XiW :

    ( ) 0.,.........2,1; NkUw NewT

    kk ==

    ],,,[ 021 NT

    X wwwW L=

    E D (W, W)

    . I E X

    , X . O, X . T . T

    .

    ()

    T DCT . T DCT

    A . (1974), 4

  • 8/10/2019 JSPTA - Real-Time Facial Recognition System-Design Implementation and Validation

    10/18

    Dr. M. Meenakshi / Journal of Signal Processing Theory and Applications (2013) 1: 1-18

    10

    W (1984). T , J

    P E G (JPEG), DCT

    1992. F , 2 DCT . T DCT

    .

    T DCT :

    1. E , F ,

    2. D ( ) ,

    / .

    T DCT N N (M , 2004; C , 1995):

    ( ) ( ) ( ) ( ) ( ) ( )

    +

    +=

    =

    = N

    vy

    N

    uxyxfvuvuF

    N

    x

    N

    y 2

    12cos2

    12cos,,

    1

    0

    1

    0

    W

    ( ) ( )

    =

    =

    0,2

    0,1

    ,

    vuN

    vuNvu

    , , , = 0, 2, 3,, N1, , (M , 2004). (x,y)

    . F(, ) DCT DCT .

    T DCT , (0, 0)

    D C (DC) . T (, )

    (, ), A

    C (AC) . I DC

    . C, DC , DCT DC (.., )

    F , ; DCT. T ,

    , . F ,

    55 DCT F. 4, DCT

    .

  • 8/10/2019 JSPTA - Real-Time Facial Recognition System-Design Implementation and Validation

    11/18

    Dr. M. Meenakshi / Journal of Signal Processing Theory and Applications (2013) 1: 1-18

    11

    0 1 1

    1 1

    1 1 1

    11 1 0

    10 1 1

    .L 5 X 5 DCT

    I DCT. F , , . F 2D DCT,

    . T,

    .

    F 5 ,

    . T

    / , :

    T :

    1) M (128 128 )

    2) O DCT 3) D DCT 2080 .

    T . T

    , , DCT DCT

    DCT128. T 128 128 DCT 64 64 . T

    64 64 128 64 1' 64 0'.

  • 8/10/2019 JSPTA - Real-Time Facial Recognition System-Design Implementation and Validation

    12/18

    Dr. M. Meenakshi / Journal of Signal Processing Theory and Applications (2013) 1: 1-18

    12

    .B DCT

    I , :

    L = T

    =

    P P B,

    12

    12 DCT M B64 6464 64T DCT

    T , :

    64

    6412 1264

    T 64 64 . A ,

    PDCT64( ) , V. T

    .

    F , V

    (S = ), :

    E = (VI S)2

  • 8/10/2019 JSPTA - Real-Time Facial Recognition System-Design Implementation and Validation

    13/18

    Dr. M. Meenakshi / Journal of Signal Processing Theory and Applications (2013) 1: 1-18

    13

    T DSP F. 6

    Start

    Make Database of Images

    Extract DCT coefficients

    512x512 input array

    Divide into 32 frames each of 16x16

    Find DCT of Each frame(2080x32 coefficients)

    Extract maximum 100coeficiients Check if match occurs

    Maximum value of ACC

    Find Autocorrelation coefficient

    Comparison

    Store this

    Extract from Matlab Extract from DSP

    Display result

    . T DCT

    .

    T .

    E MATLAB , PC DSP P . F,

    .

    T ,

    F 7 8 .

  • 8/10/2019 JSPTA - Real-Time Facial Recognition System-Design Implementation and Validation

    14/18

    Dr. M. Meenakshi / Journal of Signal Processing Theory and Applications (2013) 1: 1-18

    14

    . T P . G E

    F 9 , ..,

    E F. 10.

    .I I 10. R I

    T . T E . B

    . I E , . I E

    , .

    S MATLAB DCT . F , F. 11 512512 . F

    DCT MATLAB. O 2080 DCT , 32 .

  • 8/10/2019 JSPTA - Real-Time Facial Recognition System-Design Implementation and Validation

    15/18

    Dr. M. Meenakshi / Journal of Signal Processing Theory and Applications (2013) 1: 1-18

    15

    11.D ( 512512) 1.S

    N , , F. 12,

    . N, DCT .

    T .

    F , . T

    . T F. 13.

    1.C

    T MATLAB PC

    . B PCA DCT

    F 7 11 F 9

    12. E PC . T

    . E PCA

    DCT .

    .

    T DSP

    . T , .., . H, DSP

    TMS 320 C6713 DSP T I. TMS320C 6713 32

    200 MH.

  • 8/10/2019 JSPTA - Real-Time Facial Recognition System-Design Implementation and Validation

    16/18

    Dr. M. Meenakshi / Journal of Signal Processing Theory and Applications (2013) 1: 1-18

    16

    T , ,

    . F 14

    DSP . PCA

    DCT DSP . F

    4 . N, PCA DCT .

    1.P

    F PCA , , F. 7

    F. 9 , . E

    MATLAB PC .

    DCT , F. 11, 512512 .

    F DCT MATLAB

    DSP . S MATLAB , 32 DCT 2080

    . S F. 12

    .

    O PCA

    , . H, DSP

    , . T

    . I , . A

    , . T

    DSP. T PCA

    DSP.

    T DCT . I

    DCT , . S , MATLAB. N, DCT

    MATLAB DSP . T, DCT

    DSP. F TMS320C6713DSK

    , DSK C6713 T I.

  • 8/10/2019 JSPTA - Real-Time Facial Recognition System-Design Implementation and Validation

    17/18

    Dr. M. Meenakshi / Journal of Signal Processing Theory and Applications (2013) 1: 1-18

    17

    F PCA DCT ,

    LED , .., (

    ). A

    DSP .

    F, ,

    . I

    / .

    .

    T . B PCA DCT . A ,

    MATLAB . N, PC DSP . F

    . A , . F,

    PC .

    P. A, M, B., S, T., T, M. 1994. V

    . I P. IEEE C S. C. C V P. R., 8491.A, S., S, T., F, H., S, Y. 1991. A KL

    . I SPIE P.: I. R C V X. A

    T., 1607:7184.B, P.N., H, J.P., K, D.J., 1997. E . : R

    . IEEET. P. A. M. I., 19(7):711 720.B, R. P, T. 1993. F : F . IEEE T. P. A.

    M. I., 15(10):10421052.://../10.1109/34.254061

    B, P. 1989. M , , '' . I SPIE

    P.: V C. I P. IV., 1199:215.M. L, X. Y, K. K, W. R. K, 2004. A 2D

    DCT , P. I C I P, : 1269 1272C, R.,W, C., S, S. 1995. H : A . I P.

    IEEE, 83(5):705740.://../10.1109/5.381842

    C, I., T, D., B, A. 1992. F . I P. 2 E. C. C V,. 9296.

    D V, O. A, S., 1999,"L ". P A M I, IEEE T , 21:10, 1081 1088.

    ://../10.1109/34.799912D, R.O. H, P.E. 1973. P C S A. W: N Y, NY.

    G, V., S, S., S, D. 1990. A C . I P. 3 I. C. CV, . 718721.

  • 8/10/2019 JSPTA - Real-Time Facial Recognition System-Design Implementation and Validation

    18/18

    Dr. M. Meenakshi / Journal of Signal Processing Theory and Applications (2013) 1: 1-18

    18

    H, L.D. H, W.F., 1977. A . C G I P., 6:135 156.

    ://../10.1016/S0146664X(77)800099

    K, T. 1973. P . P.D.

    T, D I S, K U, J.K,M. 1970.V . S AI P., S, CA, T. R. AI130.

    K,M. S, L. 1990. A KL

    . IEEE T. O P. A. M I., 12:103108.://../10.1109/34.41390

    L, M.,V, J., B, J., L, J., M, C., W, R. 1993. D

    . IEEE T. C, 42:300 311.://../10.1109/12.210173

    L, K. P N., V A. N., 2003. F LDAB ", IEEE

    T N N, 14:195200.://../10.1109/TNN.2002.806647PM:18238001

    N A. V., 1999, "A H M MB A F D R", P. D.

    T, G I T. K,M. 1970.V .S AI P., S, CA, T. R. AI130.

    P. J P, H M, P R, SA R, 1997. T FERET E M F R

    A, P C C V P R, 137143. AP, A.; C, T., 2000, "F ". C, 33:2, 50 55.

    ://../10.1109/2.820039

    R. S.A, P. P.J M. H, 1998. T FERET V T P F RA, P I C A F G R.

    S, T., N, M., F, S. 1969. L . P.

    R., 1:233248.://../10.1016/00313203(69)900065

    S T, L S H A, 2009. F R M V S U

    E, I J R T E, 2:2

    S, D. W, J. 1996. U . IEEE T. P.A. M. I., 18(8):831836.

    ://../10.1109/34.531802

    R. T R S. A, 1999. T Q I C B R R WV I P C C V P R, 566

    571.

    T, M. P, A. 1991. F . I P. I. C. C. V P.R. (CVPR'91), . 586591.

    .Z. W, C. R., P .P.J., A. R, 2003. F : , ACM C.

    S. 35 (4) 399459.

    ://../10.1145/954339.954342Y G L, M.K.H., 2002. "F ". P A

    M I, IEEE T , 24:6, 764 779.

    Y A. L., C D. S. H P. W., 1992, "F E F U DT", I J C V, 8(2):99111.

    ://../10.1007/BF00127169

    Y, A., C, D., H, P. 1989. F . I P.IEEE C S. C. C V P. R., . 104109.

    Z M, H M D. L, 2001. F R U D C T,

    I J C V, 43(3):167188.

    ://../10.1023/A:1011183429707