Search

Age estimation from speech has recently received increased interest as it is useful for many applications such as user-profiling, targeted marketing, or personalized call-routing. This kind of applications need to quickly estimate the age of the speaker and might greatly benefit from real-time capabilities. Long short-term memory (LSTM) recurrent neural networks (RNN) have shown to outperform state-of-the-art approaches in related speech-based tasks, such as language identification or voice activity detection, especially when an accurate real-time response is required. In this paper, we propose a novel age estimation system based on LSTM-RNNs. This system is able to deal with short utterances (from 3 to 10 s) and it can be easily deployed in a real-time architecture. The proposed system has been tested and compared with a state-of-the-art i-vector approach using data from NIST speaker recognition evaluation 2008 and 2010 data sets. Experiments on short duration utterances show a relative improvement up to 28% in terms of mean absolute error of this new approach over the baseline system.

Automatic Age Estimation (AAE) has attracted attention due to the wide variety of possible applications. However, it is a challenging task because of the large variation of facial appearance and several other extrinsic and intrinsic factors. Most of the proposed approaches in the literature use hand-crafted features to encode ageing patterns. Deeply learned features extracted by Convolutional Neural Networks (CNNs) algorithms usually perform better than hand-crafted features. The main contribution of this paper is an extensive comparative analysis of several frameworks for real AAE based on deep learning architectures. Different well-known CNN architectures are considered and their performances are compared. MORPH, FG-NET, FACES, PubFig and CASIA-web Face datasets are used in our experiments. The robustness of the best deep estimator is evaluated under noise, expression changes, “crossing” ethnicity and “crossing” gender. The experimental results demonstrate the high performances of the popular CNNs frameworks against the state-of-art methods of automatic age estimation. A Layer-wise transfer learning evaluation is done to study the optimal number of layers to fine-tune on AAE task. An evaluation framework of Knowledge transfer from face recognition task across AAE is performed. We have made our best-performing CNNs models publicly available that would allow one to duplicate the results and for further research on the use of CNNs for AAE from face images.

The automatic classification of human images on the basis of age range and gender can be used in audiovisual content adaptation for Smart TVs or marquee advertising. Knowledge about users is used by publishing agencies and departments regulating TV content; on the basis of this information (age, gender) they are able to provide content that suits the interests of users. To this end, the creation of a highly precise image pattern recognition system is necessary, this may be one of the greatest challenges faced by computer technology in the last decades. These recognition systems must apply different pattern recognition techniques, in order to distinct gender and age in the images. In this work, we propose a multi-agent system that integrates different techniques for the acquisition, preprocessing and processing of images for the classification of age and gender. The system has been tested in an office building. Thanks to the use of a multi-agent system which allows to apply different workflows simultaneously, the performance of different methods could be compared (each flow with a different configuration). Experimental results have confirmed that a good preprocessing stage is necessary if we want the classification methods to perform well (Fisherfaces, Eigenfaces, Local Binary Patterns, Multilayer perceptron). The Fisherfaces method has proved to be more effective than MLP and the training time was shorter. In terms of the classification of age, Fisherfaces offers the best results in comparison to the rest of the system’s classifiers. The use of filters has allowed to reduce dimensionality, as a result the workload was reduced, a great advantage in a system that performs classification in real time.

The problem of time validity of biometric models has received only a marginal attention from researchers. In this paper, we propose to manage the aging influence on the adult face verification system by an A-stack age modeling technique, which uses the age as a class-independent metadata quality measure together with scores from a single or multiple baseline classifiers, in order to obtain better face verification performance. This allows for improved long-term class separation by introducing a dynamically changing decision boundary across the age progression in the scores-age space using a short-term enrollment model. This new method, based on the concept of classifier stacking and age-aware decision boundary, compares favorably with the conventional face verification approach, which uses age-independent decision threshold calculated only in the score space at the time of enrollment. Our experiments on the YouTube and MORPH data show that the use of the proposed approach allows for improving the identification accuracy as opposed to the baseline classifier.

Galbally J, Martinez-Diaz M, Fierrez J (2013) Aging in Biometrics: An Experimental Analysis on On-Line Signature. PLoS ONE 8(7): e69897. doi:10.1371/journal.pone.0069897, The first consistent and reproducible evaluation of the effect of aging on dynamic signature is reported. Experiments are carried out on a database generated from two previous datasets which were acquired, under very similar conditions, in 6 sessions distributed in a 15-month time span. Three different systems, representing the current most popular approaches in signature recognition, are used in the experiments, proving the degradation suffered by this trait with the passing of time. Several template update strategies are also studied as possible measures to reduce the impact of aging on the system’s performance. Different results regarding the way in which signatures tend to change with time, and their most and least stable features, are also given., This work has been partially supported by projects Contexts (S2009/TIC-1485) from CAM, by Bio-Challenge (TEC2009-11186) and Bio-Shield (TEC2012-34881) from Spanish MECD, by Cátedra UAM-Telefónica, and by the Spanish Dirección General de la Guardia Civil.

The problem of time validity of biometric models has received only a marginal attention from researchers. Actual and up-to-date at the time of their creation, extracted features and models relevant to a person's face may eventually become outdated, leading to a failure in the face identification task. If physical characteristics of the individual change over time, their classification model has to be updated. In this paper we present a mutli-classifier A-stack scheme, which is based on the concept of classifier stacking and makes use of the age information and scores of multiple baseline classifiers, in order to improve the identification performance during age progression. Our experiments on the MORPH database show that the use of the proposed multi-classifier stacking fusion allows for improving the identification accuracy as opposed to the baseline classifier and single-classifier A-stack method.

Phone-like acoustic models (AMs) used in large-vocabulary automatic speech recognition (ASR) systems are usually trained with speech collected from young adult speakers. Using such models, ASR performance may decrease by about 10% absolute when transcribing elderly speech. Ageing is known to alter speech production in ways that require ASR systems to be adapted, in particular at the level of acoustic modeling. In this study, we investigated automatic age estimation in order to select age-specific adapted AMs. A large corpus of read speech from European Portuguese speakers aged 60 or over was used. Age estimation (AE) based on i-vectors and support vector regression achieved mean error rates of about 4.2 and 4.5 years for males and females, respectively. Compared with a baseline ASR system with AMs trained using young adult speech and a WER of 13.9%, the selection of five-year-range adapted AMs, based on the estimated age of the speakers, led to a decrease in WER of about 9.3% relative (1.3% absolute). Comparable gains in ASR performance were observed when considering two larger age ranges (60-75 and 76-90) instead of six five-year ranges, suggesting that it would be sufficient to use the two large ranges only. info:eu-repo/semantics/acceptedVersion