Lieuwe Jan Spreeuwers

Fast and Accurate 3D Face Recognition

In this paper we present a new robust approach for 3D face registration to an intrinsic coordinate system of the face. The intrinsic coordinate system is defined by the vertical symmetry plane through the nose, the tip of the nose and the slope of the bridge of the nose. In addition, we propose a 3D face classifier based on the fusion of many dependent region classifiers for overlapping face regions. The region classifiers use PCA-LDA for feature extraction and the likelihood ratio as a matching score. Fusion is realised using straightforward majority voting for the identification scenario. For verification, a voting approach is used as well and the decision is defined by comparing the number of votes to a threshold. Using the proposed registration method combined with a classifier consisting of 60 fused region classifiers we obtain a 99.0% identification rate on the all vs first identification test of the FRGC v2 data. Axa0verification rate of 94.6% at FAR=0.1% was obtained for the all vs all verification test on the FRGC v2 data using fusion of 120 region classifiers. The first is the highest reported performance and the second is in the top-5 of best performing systems on these tests. In addition, our approach is much faster than other methods, taking only 2.5 seconds per image for registration and less than 0.1 ms per comparison. Because we apply feature extraction using PCA and LDA, the resulting template size is also very small: 6xa0kB for 60 region classifiers.

The Effect of Image Resolution on the Performance of a Face Recognition System

In this paper we investigate the effect of image resolution on the error rates of a face verification system. We do not restrict ourselves to the face recognition algorithm only, but we also consider the face registration. In our face recognition system, the face registration is done by finding landmarks in a face image and subsequent alignment based on these landmarks. To investigate the effect of image resolution we performed experiments where we varied the resolution. We investigate the effect of the resolution on the face recognition part, the registration part and the entire system. This research also confirms that accurate registration is of vital importance to the performance of the face recognition algorithm. The results of our face recognition system are optimal on face images with a resolution of 32 times 32 pixels

Robust 3D face recognition in the presence of realistic occlusions

Facial occlusions pose significant problems for automatic face recognition systems. In this work, we propose a novel occlusion-resistant three-dimensional (3D) facial identification system. We show that, under extreme occlusions due to hair, hands, and eyeglasses, typical 3D face recognition systems exhibit poor performance. In order to deal with occlusions, our proposed system employs occlusion-resistant registration, occlusion detection, and regional classifiers. A two-step registration module first detects the nose region on the curvedness-weighted convex shape index map, and then performs fine alignment using nose-based Iterative Closest Point (ICP) algorithm. Occluded areas are determined automatically via a generic face model. After non-facial parts introduced by occlusions are removed, a variant of Gappy Principal Component Analysis (Gappy PCA) is used to restore the full face from occlusion-free facial surfaces. Experimental results obtained on realistically occluded facial images from the Bosphorus 3D face database shows that, with the use of score-level fusion of regional Linear Discriminant Analysis (LDA) classifiers, the proposed method improves rank-1 identification accuracy significantly: from 76.12% to 94.23%.

A Bootstrap Approach to Eigenvalue Correction

Eigenvalue analysis is an important aspect in many data modeling methods. Unfortunately, the eigenvalues of the sample covariance matrix (sample eigenvalues) are biased estimates of the eigenvalues of the covariance matrix of the data generating process (population eigenvalues). We present a new method based on bootstrapping to reduce the bias in the sample eigenvalues: the eigenvalue estimates are updated in several iterations, where in each iteration synthetic data is generated to determine how to update the population eigenvalue estimates. Comparison of the bootstrap eigenvalue correction with a state of the art correction method by Karoui shows that depending on the type of population eigenvalue distribution, sometimes the Karoui method performs better and sometimes our bootstrap method.

A neural network edge detector

Extracting edges from images is a widely used first step in processing. A different view on the well known enhancement/thresholding approach for edge detection is presented in this paper. The structure of a two layer feed forward neural network is comparable to the structure of enhancement/thresholding edge detectors. It is possible to calculate an optimal edge detector with a certain predefined network structure and training set, by training the neural network with examples of edge and nonedge patterns. The back propagation learning rule is used for optimization of the network. The choice of the network structure and the training set are very important, because they determine the final behavior of the network. The paper describes which network structures were selected and how the training sets were generated. Some of the experiments are described, and observations of the convolution kernels for edge enhancement that are formed during training. Finally the results are evaluated and compared with the results of edge detectors based on the Sobel, Marr-Hildreth and Canny edge enhancement algorithms. It appears that the neural network edge detector can be made very robust against noise and blur and in most tests outperforms the others.

Towards Automatic Forensic Face Recognition

In this paper we present a methodology and experimental results for evidence evaluation in the context of forensic face recognition. In forensic applications, the matching score (hereafter referred to as similarity score) from a biometric system must be represented as a Likelihood Ratio (LR). In our experiments we consider the face recognition system as a ‘black box’ and compute LR from similarity scores. The proposed approach is in accordance with the Bayesian framework where the duty of a forensic scientist is to compute LR from biometric evidence which is then incorporated with prior knowledge of the case by the judge or jury. In our experiments we use a total of 2878 images of 100 subjects from two different databases. Our experimental results prove the feasibility of our approach to reach a LR value given an image of a suspect face and questioned face. In addition, we compare the performance of two biometric face recognition systems in forensic casework.

Effect of calibration data on forensic likelihood ratio from a face recognition system

A biometric system used for forensic evaluation requires a conversion of the score to a likelihood ratio. A likelihood ratio can be computed as the ratio of the probability of a score given the prosecution hypothesis is true and the probability of a score given the defense hypothesis is true. In this paper we study two different approaches of a forensic likelihood ratio computation in the context of forensic face recognition. These approaches differ in the databases they use to obtain the score distribution under the prosecution and the defense hypothesis and therefore consider slightly different interpretation of these hypotheses. The goal of this study is to quantify the effect of these approaches on the resultant likelihood ratio in the context of evidence evaluation from a face recognition system. A state-of-the art commercial face recognition system is employed for facial images comparison and computation of scores. A simple forensic case is simulated by randomly selecting a small subset from the FRGC database. Images in this subset are used to estimate the score distribution under the prosecution and the defense hypothesis and the effect of different approaches of a likelihood ratio computation is demonstrated and explained. It is observed that there is a significant variation in the resultant likelihood ratios given the databases which are used to model the prosecution and defense hypothesis are varied.

Virtual illumination grid for correction of uncontrolled illumination in facial images

Face recognition under uncontrolled illumination conditions is still considered an unsolved problem. In order to correct for these illumination conditions, we propose a virtual illumination grid (VIG) approach to model the unknown illumination conditions. Furthermore, we use coupled subspace models of both the facial surface and albedo to estimate the face shape. In order to obtain a representation of the face under frontal illumination, we relight the estimated face shape. We show that the frontal illuminated facial images achieve better performance in face recognition. We have performed the challenging Experiment 4 of the FRGCv2 database, which compares uncontrolled probe images to controlled gallery images. Our illumination correction method results in considerably better recognition rates for a number of well-known face recognition methods. By fusing our global illumination correction method with a local illumination correction method, further improvements are achieved.

A Study of Identification Performance of Facial Regions from CCTV Images

This paper focuses on automatic face identification for forensic applications. Forensic examiners compare different parts of the face image obtained from a closed-circuit television (CCTV) image with a database of mug shots or good quality image(s) taken from the suspect. In this work we study and compare the discriminative capabilities of different facial regions (also referred to as facial features) such as eye, eyebrow, mouth, etc. It is useful because it can statistically support the current practice of forensic facial comparison. It is also of interest to biometrics as a more robust general-purpose face recognition system can be built by fusing the similarity scores obtained from the comparison of different individual parts of the face. For experiments with automatic systems, we simulate a very challenging recognition scenario by using a database of 130 subjects each having only one gallery image. Gallery images are frontal mug shots while probe set consist of low quality CCTV camera images. Face images in gallery and probe sets are first segmented using eye locations and recognition experiments are performed for the different face regions considered. We also study and evaluate an improved recognition approach based on AdaBoost algorithm with Linear Discriminant Analysis (LDA) as a week learner and compare its performance with the baseline Eigenface method for automatic facial feature recognition.

Breaking the 99% barrier: optimisation of 3D face recognition

This study presents optimisations to a three-dimensional (3D) face recognition method the authors published in 2011. The optimisations concern handling and estimation of motion from a single 3D image using the symmetry of the face, fine registration by selection of the maximum score for small variations of the registration parameters and efficient training using automatic outlier removal where only part of the classifier is retrained. The optimisations lead to a staggering performance improvement: the verification rate on Face Recognition Grand Challenge (FRGC) v2 data at false accept rate = 0.1% increases from 94.6 to 99.3% and the identification rate increases from 99 to 99.4%. Both are, to the authors knowledge, the best scores ever published on the FRGC data. In addition, the registration time was reduced from about 2.5 to 0.2–0.6 s and the number of comparisons has increased from about 11 000 to more than 50 000 per second. For slightly decreased performance, even millions of comparisons can be realised. The fast registration means near real-time recognition with 2–5 images is possible. The optimisations are not specific for this method, but can be beneficial for other 3D face recognition methods as well.