Benoît Duc

Expert Conciliation for Multi Modal Person Authentication Systems by Bayesian Statistics

We present an algorithm functioning as a supervisor module in a multi expert decision making machine. It uses the Bayes theory in order to estimate the biases of individual expert opinions. These are then used to calibrate and conciliate expert opinions to one opinion. We present a framework for simulating decision strategies using expert opinions whose properties are easily modifiable. By using real data coming from a person authentication system using image and speech data we were able to confirm that the proposed supervisor improves the quality of individual expert decisions by reaching success rates of 99.5 %.

Fusion of audio and video information for multi modal person authentication

Abstract We present an algorithm functioning as a supervisor module in a multi-expert decision making machine. It uses the Bayes theory in order to estimate the biases of individual expert opinions. The biases are used to calibrate and conciliate expert opinions to a single decision. This supervision technique is applied to the real case of a person authentication technique using two modalities, face and speech. The visual part involves the matching of a coarse grid containing Gabor phase information from face images. The acoustic part is performed by a text-dependent speaker verification system based on Hidden Markov Models. Experimental results show that the proposed fusion method improves the quality of individual expert decisions by reaching success rates of 99.5%.

Person Authentication by Fusing Face and Speech Information

Keywords: vision Reference EPFL-CONF-82406 Record created on 2006-03-10, modified on 2017-05-10

Multi-Modal Person Authentication

This paper deals with the elements of a multi-modal person authentication systems. Test procedures for evaluating machine experts as well as machine supervisors based on leave-one-out principle are described. Two independent machine experts on person authentication are presented along with their individual performances. These experts consisted of a face (Gabor features) and a speaker (LPC features) authentication algorithm trained on the M2VTS multi-media database. The expert opinions are combined yielding far better performances by using a trained supervisor based on Bayesian statistics than individual modalities aggregated by averaging.

Face authentication with sparse grid Gabor information

This paper investigates the application of statistical pattern recognition methods in the framework of the dynamic link matching approach. This method describes objects by means of local frequency information on nodes of a sparse grid. Matching of an input image with a reference is achieved by displacement and deformation of the grid. This method is applied here to the authentication of human faces in a cooperative scenario where candidates claim an identity that is to be checked. The matching error is not powerful enough to provide satisfying results in this case. We introduce an automatic weighting of the nodes according to their significance. Results show that for regular grids, this weighting leads to a significant improvement of the performance.

Spatio-Temporal Robust Motion Estimation and Segmentation

In this paper, a general spatio-temporal framework for motion estimation is presented. It allows to estimate a fully parametric motion model over an image sequence. As parametric models describe one motion only, a robust estimator is introduced in order to cope with several moving objects. The motion segmentation algorithm combines luminance and the composition of all the motions detected over a set of successive frames for motion boundaries estimation.

Face Authentication Using Morphological Dynamic Link Architecture

A very attractive approach for face detection is based on multiresolution images (also known as mosaic images). Motivated by the simplicity of this approach, a rule-based face detection algorithm in frontal views is developed first. Second, a novel dynamic link architecture based on multiscale morphological dilation-erosion is proposed for face authentication. More specifically, a sparse grid is placed over the outcome of face detection stage for each person in a reference set. Subsequently, multiscale morphological operations are employed to yield a feature vector at each node of the grid and dynamic link matching is applied to verify the identity of each person from a test set. The first experimental results reported in this paper verify the superiority of the proposed method over the (standard) dynamic link matching that is based on Gabor wavelets.

Shape normalisation for face recognition

This paper presents methods for shape normalisation of face images. Localisation and shape normalisation are prerequisites for face recognition algorithms like the eigenface approach. Other established face recognition methods like labeled graph matching can also be accelerated by providing normalised face image databases.

Motion estimation using invariance under group transformations

This paper describes a spatiotemporal framework for motion estimation, based on a fully parametric motion model. The model type defines a group of transformations. The motion estimation consists in finding a subgroup which leaves the image sequence invariant.

Motion segmentation by fuzzy clustering with automatic determination of the number of motions

A layered motion estimation scheme using fuzzy clustering is introduced in this paper. Once motion estimation is performed a modified objective criterion is applied to discard non significant classes.