SNL3dFace
Abstract
This software distribution contains MATLAB and C++ code to enable identity verification using 3D images that may or may not contain a texture component. The code is organized to support system performance testing and system capability demonstration through the proper configuration of the available user interface. Using specific algorithm parameters the face recognition system has been demonstrated to achieve a 96.6% verification rate (Pd) at 0.001 false alarm rate. The system computes robust facial features of a 3D normalized face using Principal Component Analysis (PCA) and Fisher Linear Discriminant Analysis (FLDA). A 3D normalized face is obtained by alighning each face, represented by a set of XYZ coordinated, to a scaled reference face using the Iterative Closest Point (ICP) algorithm. The scaled reference face is then deformed to the input face using an iterative framework with parameters that control the deformed surface regulation an rate of deformation. A variety of options are available to control the information that is encoded by the PCA. Such options include the XYZ coordinates, the difference of each XYZ coordinates from the reference, the Z coordinate, the intensity/texture values, etc. In addition to PCA/FLDA feature projection this software supports feature matching to obtain similarity matricesmore »
- Authors:
- Publication Date:
- Research Org.:
- Sandia National Laboratories (SNL), Albuquerque, NM, and Livermore, CA (United States)
- Sponsoring Org.:
- Sandia National Laboratories
- OSTI Identifier:
- 1231087
- Report Number(s):
- SNL3dFace; 002240IBMPC00
R&D Project: Secure Portal LDRD 93590
- Resource Type:
- Software
- Software Revision:
- 00
- Software Package Number:
- 002240
- Software Package Contents:
- Media Directory; Software Abstract; Media includes Source Code; Object Library; Compilation Instructions; Linking Instructions; Executable Module(s); User Guide / 1 CD-ROM
- Software CPU:
- IBMPC
- Open Source:
- No
- Source Code Available:
- Yes
- Related Software:
- Uses Visual Toolkit (VTK) from www.vtk.org
- Country of Publication:
- United States
Citation Formats
Russ, Trina, Koch, Mark, Koudelka, Melissa, Peters, Ralph, Little, Charles, Boehnen, Chris, and Peters, Tanya. SNL3dFace.
Computer software. Vers. 00. Sandia National Laboratories. 20 Jul. 2007.
Web.
Russ, Trina, Koch, Mark, Koudelka, Melissa, Peters, Ralph, Little, Charles, Boehnen, Chris, & Peters, Tanya. (2007, July 20). SNL3dFace (Version 00) [Computer software].
Russ, Trina, Koch, Mark, Koudelka, Melissa, Peters, Ralph, Little, Charles, Boehnen, Chris, and Peters, Tanya. SNL3dFace.
Computer software. Version 00. July 20, 2007.
@misc{osti_1231087,
title = {SNL3dFace, Version 00},
author = {Russ, Trina and Koch, Mark and Koudelka, Melissa and Peters, Ralph and Little, Charles and Boehnen, Chris and Peters, Tanya},
abstractNote = {This software distribution contains MATLAB and C++ code to enable identity verification using 3D images that may or may not contain a texture component. The code is organized to support system performance testing and system capability demonstration through the proper configuration of the available user interface. Using specific algorithm parameters the face recognition system has been demonstrated to achieve a 96.6% verification rate (Pd) at 0.001 false alarm rate. The system computes robust facial features of a 3D normalized face using Principal Component Analysis (PCA) and Fisher Linear Discriminant Analysis (FLDA). A 3D normalized face is obtained by alighning each face, represented by a set of XYZ coordinated, to a scaled reference face using the Iterative Closest Point (ICP) algorithm. The scaled reference face is then deformed to the input face using an iterative framework with parameters that control the deformed surface regulation an rate of deformation. A variety of options are available to control the information that is encoded by the PCA. Such options include the XYZ coordinates, the difference of each XYZ coordinates from the reference, the Z coordinate, the intensity/texture values, etc. In addition to PCA/FLDA feature projection this software supports feature matching to obtain similarity matrices for performance analysis. In addition, this software supports visualization of the STL, MRD, 2D normalized, and PCA synthetic representations in a 3D environment.},
doi = {},
url = {https://www.osti.gov/biblio/1231087},
year = {Fri Jul 20 00:00:00 EDT 2007},
month = {Fri Jul 20 00:00:00 EDT 2007},
note =
}
To receive further information, fill out the request form below. OSTI staff will begin to process an order for scientific and technical software once the signed site license agreement is received. You may also reach us by email at: .