José Luis Alba-Castro

Toward Pose-Invariant 2-D Face Recognition Through Point Distribution Models and Facial Symmetry

This paper proposes novel ways to deal with pose variations in a 2-D face recognition scenario. Using a training set of sparse face meshes, we built a point distribution model and identified the parameters which are responsible for controlling the apparent changes in shape due to turning and nodding the head, namely the pose parameters. Based on them, we propose two approaches for pose correction: 1) a method in which the pose parameters from both meshes are set to typical values of frontal faces, and 2) a method in which one mesh adopts the pose parameters of the other one. Finally, we obtain pose corrected meshes and, taking advantage of facial symmetry, virtual views are synthesized via Thin Plate Splines-based warping. Given that the corrected images are not embedded into a constant reference frame, holistic methods are not suitable for feature extraction. Instead, the virtual faces are fed into a system that makes use of Gabor filtering for recognition. Unlike other approaches that warp faces onto a mean shape, we show that if only pose parameters are modified, client specific information remains in the warped image and discrimination between subjects is more reliable. Statistical analysis of the authentication results obtained on the XM2VTS database confirm the hypothesis. Also, the CMU PIE database is used to assess the performance of the proposed methods in an identification scenario where large pose variations are present, achieving state-of-the-art results and outperforming both research and commercial techniques.

Single- and cross- database benchmarks for gender classification under unconstrained settings

Gender classification is one of the most important tasks in automated face analysis, and has attracted the interest of researchers for years. Up to now, most gender classification approaches have been tested using single-database experiments, and on quite controlled datasets such as the FERET database, which are not representative of real world settings. However, a recent trend towards more realistic benchmarks has emerged within the face analysis community, leading to the appearance of databases and protocols such as the Labeled Faces in the Wild (LFW) database, and the so-called Gallaghers database, which comprises images collected from Flickr.

On combining classifiers for speaker authentication

Abstract Speaker verification and utterance verification are examples of techniques that can be used for speaker authentication purposes. Speaker verification consists of accepting or rejecting the claimed identity of a speaker by processing samples of his/her voice. Usually, these systems are based on HMM models that try to represent the characteristics of the speakers’ vocal tracts. Utterance verification systems make use of a set of speaker-independent speech models to recognize a certain utterance. If the utterances consist of passwords, this can be used for identity verification purposes. Up to now, both techniques have been used separately. This paper is focused on the problem of how to combine these two sources of information. New architectures are presented to join an utterance verification system and a speaker verification system in order to improve the performance in a speaker verification task.

Shape-Driven Gabor Jets for Face Description and Authentication

This paper proposes, through the combination of concepts and tools from different fields within the computer vision community, an alternative path to the selection of key points in face images. The classical way of attempting to solve the face recognition problem using algorithms which encode local information is to localize a predefined set of points in the image, extract features from the regions surrounding those locations, and choose a measure of similarity (or distance) between correspondent features. Our approach, namely shape-driven Gabor jets, aims at selecting an own set of points and features for a given client. After applying a ridges and valleys detector to a face image, characteristic lines are extracted and a set of points is automatically sampled from these lines where Gabor features (jets) are calculated. So each face is depicted by R2 points and their respective jets. Once two sets of points from face images have been extracted, a shape-matching algorithm is used to solve the correspondence problem (i.e., map each point from the first image to a point within the second image) so that the system is able to compare shape-matched jets. As a byproduct of the matching process, geometrical measures are computed and compiled into the final dissimilarity function. Experiments on the AR face database confirm good performance of the method against expression and, mainly, lighting changes. Moreover, results on the XM2VTS and BANCA databases show that our algorithm achieves better performance than implementations of the elastic bunch graph matching approach and other related techniques.

Shedding light on the asymmetric learning capability of AdaBoost

In this paper, we propose a different insight to analyze AdaBoost. This analysis reveals that, beyond some preconceptions, AdaBoost can be directly used as an asymmetric learning algorithm, preserving all its theoretical properties. A novel class-conditional description of AdaBoost, which models the actual asymmetric behavior of the algorithm, is presented.

Modeling Gabor Coefficients via Generalized Gaussian Distributions for Face Recognition

Gabor filters are biologically motivated convolution kernels that have been widely used in the field of computer vision and, specially, in face recognition during the last decade. This paper proposes a statistical model of Gabor coefficients extracted from face images using generalized Gaussian distributions (GGDs). By measuring the Kullback-Leibler distance (KLD) between the pdf of the GGD and the relative frequency of the coefficients, we conclude that GGDs provide an accurate modeling. The underlying statistics allow us to reduce the required amount of data to be stored (i.e. data compression) via Lloyd-Max quantization. Verification experiments on the XM2VTS database show that performance does not drop when, instead of the original data, we use quantized coefficients.

What Your Face Vlogs About: Expressions of Emotion and Big-Five Traits Impressions in YouTube

Social video sites where people share their opinions and feelings are increasing in popularity. The face is known to reveal important aspects of human psychological traits, so the understanding of how facial expressions relate to personal constructs is a relevant problem in social media. We present a study of the connections between automatically extracted facial expressions of emotion and impressions of Big-Five personality traits in YouTube vlogs (i.e., video blogs). We use the Computer Expression Recognition Toolbox (CERT) system to characterize users of conversational vlogs. From CERT temporal signals corresponding to instantaneously recognized facial expression categories, we propose and derive four sets of behavioral cues that characterize face statistics and dynamics in a compact way. The cue sets are first used in a correlation analysis to assess the relevance of each facial expression of emotion with respect to Big-Five impressions obtained from crowd-observers watching vlogs, and also as features for automatic personality impression prediction. Using a dataset of 281 vloggers, the study shows that while multiple facial expression cues have significant correlation with several of the Big-Five traits, they are only able to significantly predict Extraversion impressions with moderate values of R2.

Local Contour Patterns for fast traffic sign detection

Advanced driver assistance systems have strong restrictions for real-time performance. Vision algorithms embedded in these systems need to balance accuracy and computational simplicity and there exists a continuous challenge to increase both goals. In this paper we define a new operator coined as Local Contour Patterns and use it in fast Hough-Transform-based approaches for circle and line detectors. We show an efficient implementation for traffic sign detection, relying only on shape information, that analyzes a 752×480 grayscale image in 40 ms in a Intel 8400 CPU, with a very good performance in real driving conditions.

Double-base asymmetric AdaBoost

Based on the use of different exponential bases to define class-dependent error bounds, a new and highly efficient asymmetric boosting scheme, coined as AdaBoostDB (Double-Base), is proposed. Supported by a fully theoretical derivation procedure, unlike most of the other approaches in the literature, our algorithm preserves all the formal guarantees and properties of original (cost-insensitive) AdaBoost, similarly to the state-of-the-art Cost-Sensitive AdaBoost algorithm. However, the key advantage of AdaBoostDB is that our novel derivation scheme enables an extremely efficient conditional search procedure, dramatically improving and simplifying the training phase of the algorithm. Experiments, both over synthetic and real datasets, reveal that AdaBoostDB is able to save over 99% training time with regard to Cost-Sensitive AdaBoost, providing the same cost-sensitive results. This computational advantage of AdaBoostDB can make a difference in problems managing huge pools of weak classifiers in which boosting techniques are commonly used.

Fast real-time multiclass traffic sign detection based on novel shape and texture descriptors

Detection and classification of traffic signs is one of the most studied Advanced Driver Assistance Systems (ADAS) and some solutions are already installed in vehicles. Nevertheless these systems still have room for improvement in terms of speed and performance. When driving at high speed, warning systems require very fast processing of the video stream in order to lose as few frames as possible and minimize the chance of missing a readable traffic sign. In this paper we show a sign detection system for grayscale images based on a two-stage process: A rapid shape prefiltering, that relies on a new descriptor coined as Local Contour Patterns, rejects most of the image subwindows and preclassifies the rest as one of the three main sign types. Then, a sign-dependent AdaBoost-based cascade detector that makes use of a new set of simpler texture features, coined as Quantum Features, scans the pre-fetched subwindows to fine tune candidate traffic signs. The analysis of this detector over hundreds of video sequences which were captured with a car-mounted 752×480 grayscale camera and provided by the Galician Automotive Technology Center (CTAG) shows a very good behavior for multiclass traffic sign detection running at 14 frames/sec on a 2.8 GHz processor.