3dfacerec-combinedspaces

Upload: prabhu-teja

Post on 03-Jun-2018

214 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/12/2019 3DFaceRec-CombinedSpaces

    1/14

    Three-Dimensional Face Recognition UsingSurface Space Combinations

    Thomas Heseltine, Nick Pears, Jim AustinAdvanced Computer Architecture Group

    Department of Computer Science - University of York

    www.cs.york.ac.uk/~tomh [email protected]

  • 8/12/2019 3DFaceRec-CombinedSpaces

    2/14

    2

    Introduction

    Face recognition offers several advantages over other biometrics

    Covert operation.

    Human readable media.

    Public acceptance.

    Data required is readily availablepolice databases etc.

    But

    Growing interest in biometric authentication

    National ID cards, Airport security (MRPs), Surveillance.

    Fingerprint, iris, hand geometry, gait, voice, vein and face.

  • 8/12/2019 3DFaceRec-CombinedSpaces

    3/14

    3

    Limitations of 2D Face Recognition

    Lighting conditions.

    Different lighting conditions for enrolmentand query.

    Bright light causing image saturation.

    Head orientation.2D feature distances appear to distort.

    Image quality.

    CCTV, Web-cams etc.

    Facial expression.

    Changes in feature location and shape.

    Partial occlusion

    Hats, scarves, glasses etc.

    System effectiveness is highly dependant on image capture conditions.

    Face recognition is not as accurate as other biometrics.

    Error rates that are too high for many applications in mind.Result:

  • 8/12/2019 3DFaceRec-CombinedSpaces

    4/14

  • 8/12/2019 3DFaceRec-CombinedSpaces

    5/14

    5

    3D Face DataGenerated using a stereo vision cameraenhanced by light projection.

    Stored in OBJ file format.

    Approximately 8000 points on a facial surface.

    Greyscale texture mapped.

    Wire-mesh TexturePolygons

    Lighting

  • 8/12/2019 3DFaceRec-CombinedSpaces

    6/14

    6

    The Fishersurface Method

    Developed in previous work[Heseltine, Pears, Austin. Three-Dimensional Face Recognition: A Fishersurface Approach].

    Adaptation of the fisherface method to 3D face data.

    [Belhumeur, Hespanha, Kriegman, Eigenfaces vs. Fisherfaces: Face Recognition using class specific linear projection].

    Uses PCA + LDA to create a surface space projection matrix

    Orientate 3D face models to face directly forwards.

    Convert to depth-map representation (60 by 90 pixels).

    Train on 300 depth maps of 50 different people.

    Projected depth maps compared using Euclidean or cosine

    distance metrics.

  • 8/12/2019 3DFaceRec-CombinedSpaces

    7/14

    7

  • 8/12/2019 3DFaceRec-CombinedSpaces

    8/14

    8

    Test DatabaseLittle publicly available 3D Face data, so we collect our own 3D face database:

    Database now consists of over 5000 face models of over 350 people.

    Large range of expression, orientation, gender, ethnicity, age.

    We take a subset of this database (1770 models) for training and testing.

    300 3D models of 50 people for training

    1470 3D models of 280 people for testing

  • 8/12/2019 3DFaceRec-CombinedSpaces

    9/14

    9

    Error Rates

    Error curves produced for all surfacerepresentations.

    EER taken as a single comparative

    value.

    A large range of error rates produced.

  • 8/12/2019 3DFaceRec-CombinedSpaces

    10/14

    10

    Surface Space Analysis Using FLD

    c

    i x

    i

    c

    i

    i

    i

    i

    mx

    mm

    d

    1

    21

    1

    2

    )(

    )( Fishers Linear Discriminant calculates the ratio of

    between-class and within-class scatter, providing

    an indication of discriminating ability.

  • 8/12/2019 3DFaceRec-CombinedSpaces

    11/14

    11

    Combining Surface Space Dimensions

    However, even the worst representations

    produce a surface space with some highly

    discriminatory dimensions.

    Some surface representationsperform better than others.

    Extract best dimensions

    from all surface spaces

    Incorporate into a single

    combined surface space

    Dividing each element by its within-class standard deviation

    effectively weights each dimension evenly.

  • 8/12/2019 3DFaceRec-CombinedSpaces

    12/14

    12

    Test Procedure

  • 8/12/2019 3DFaceRec-CombinedSpaces

    13/14

  • 8/12/2019 3DFaceRec-CombinedSpaces

    14/14

    Questions?

    Thomas Heseltine, Nick Pears, Jim Austin

    Advanced Computer Architecture Group

    Department of Computer Science - University of York

    www.cs.york.ac.uk/~tomh [email protected]

    Three-Dimensional Face Recognition Using

    Surface Space Combinations