Face authentication with Gabor information on deformable graphs

Abstract

Abstract-Elastic graph matching has been proposed as a practical implementation of dynamic link matching, which is a neural network with dynamically evolving links between a reference model and an input image. Each node of the graph contains features that characterize the neighborhood of its location in the image. The elastic graph matching usually consists of two consecutive steps, namely a matching with a rigid grid, followed by a deformation of the grid, which is actually the elastic part. The deformation step is introduced in order to allow for some deformation, rotation, and scaling of the object to be matched. This method is applied here to the authentication of human faces where candidates claim an identity that is to be checked. The matching error as originally suggested is not powerful enough to provide satisfying results in this case. We introduce an automatic weighting of the nodes according to their significance. We also explore the significance of the elastic deformation for an application of face-based person authentication. We compare performance results obtained with and without the second matching step. Results show that the deformation step slightly increases the performance, but has lower influence than the weighting of the nodes. The best results are obtained with the combination of both aspects. The results provided by the proposed method compare favorably with two methods that require a prior geometric face normalization, namely the synergetic and eigenface approaches.

Robust Features for Frontal Face Authentication in Difficult Image Conditions

“…In this work we shall concentrate on the feature extraction part of the last stage. There are many approaches to facial feature extraction; for example, Turk and Pentland [17] used Principal Component Analysis (PCA), Duc et al [5] used biologically inspired 2D Gabor wavelets, while Eickeler et al [7] obtained features using the 2D Discrete Cosine Transform (DCT). Recently, Sanderson & Paliwal [14] used a modified form of DCT feature extraction, termed DCT-mod2, which has been shown to be robust against illumination direction changes.…”

Exploiting discriminant information in elastic graph matching

“…A popular class of techniques used for frontal face recognition/verification is EGM [1]. In EGM the reference object graph is created by projecting the object's image onto a rectangular elastic sparse graph where a Gabor wavelet bank response is measured at each node.…”

“…In EGM the reference object graph is created by projecting the object's image onto a rectangular elastic sparse graph where a Gabor wavelet bank response is measured at each node. The graph matching procedure is implemented by a coarse-to-fine stochastic optimization of a cost function which takes into account both jet similarities and node deformation [1].…”

On Accuracy/Robustness/Complexity Trade-Offs in Face Verification

“…number of FAs number of impostor accesses (6) FRR = number of FRs number of true claimant accesses (7) To aid the interpretation of accuracy, the two error measures are often combined using the Half Total Error Rate (HTER), defined as [2]:…”

“…The use of the face as a biometric is particularly attractive, as it can involve little or no interaction with the person to be verified [15]. Various techniques have been proposed for face classification; some examples are systems based on Principal Component Analysis (PCA) feature extraction [24], modular PCA [16], Elastic Graph Matching (EGM) [6], and Support Vector Machines [19]. Examples specific to statistical models include one-dimensional Hidden Markov Models (1D HMMs) [20], pseudo-2D HMMs [7] and Gaussian Mixture Models (GMMs) [3], [21] (which can be considered as a simplified version of HMMs).…”