Ezequiel López-Rubio

Restoration of images corrupted by Gaussian and uniform impulsive noise

Many approaches to image restoration are aimed at removing either Gaussian or uniform impulsive noise. This is because both types of degradation processes are distinct in nature, and hence they are easier to manage when considered separately. Nevertheless, it is possible to find them operating on the same image, which produces a hard damage. This happens when an image, already contaminated by Gaussian noise in the image acquisition procedure, undergoes impulsive corruption during its digital transmission. Here we propose a principled method to remove both types of noise. It is based on a Bayesian classification of the input pixels, which is combined with the kernel regression framework.

FOREGROUND DETECTION IN VIDEO SEQUENCES WITH PROBABILISTIC SELF-ORGANIZING MAPS

Background modeling and foreground detection are key parts of any computer vision system. These problems have been addressed in literature with several probabilistic approaches based on mixture models. Here we propose a new kind of probabilistic background models which is based on probabilistic self-organising maps. This way, the background pixels are modeled with more flexibility. On the other hand, a statistical correlation measure is used to test the similarity among nearby pixels, so as to enhance the detection performance by providing a feedback to the process. Several well known benchmark videos have been used to assess the relative performance of our proposal with respect to traditional neural and non neural based methods, with favourable results, both qualitatively and quantitatively. A statistical analysis of the differences among methods demonstrates that our method is significantly better than its competitors. This way, a strong alternative to classical methods is presented.

A principal components analysis self-organizing map

We propose a new self-organizing neural model that performs principal components analysis. It is also related to the adaptive subspace self-organizing map (ASSOM) network, but its training equations are simpler. Experimental results are reported, which show that the new model has better performance than the ASSOM network.

Kernel regression based feature extraction for 3D MR image denoising

Kernel regression is a non-parametric estimation technique which has been successfully applied to image denoising and enhancement in recent times. Magnetic resonance 3D image denoising has two features that distinguish it from other typical image denoising applications, namely the tridimensional structure of the images and the nature of the noise, which is Rician rather than Gaussian or impulsive. Here we propose a principled way to adapt the general kernel regression framework to this particular problem. Our noise removal system is rooted on a zeroth order 3D kernel regression, which computes a weighted average of the pixels over a regression window. We propose to obtain the weights from the similarities among small sized feature vectors associated to each pixel. In turn, these features come from a second order 3D kernel regression estimation of the original image values and gradient vectors. By considering directional information in the weight computation, this approach substantially enhances the performance of the filter. Moreover, Rician noise level is automatically estimated without any need of human intervention, i.e. our method is fully automated. Experimental results over synthetic and real images demonstrate that our proposal achieves good performance with respect to the other MRI denoising filters being compared.

Assessment of geometric features for individual identification and verification in biometric hand systems

This paper studies the reliability of geometric features for the identification of users based on hand biometrics. Our methodology is based on genetic algorithms and mutual information. The aim is to provide a system for user identification rather than a classification. Additionally, a robust hand segmentation method to extract the hand silhouette and a set of geometric features in hard and complex environments is described. This paper focuses on studying how important and discriminating the hand geometric features are, and if they are suitable in developing a robust and reliable biometric identification. Several public databases have been used to test our method. As a result, the number of required features have been drastically reduced from datasets with more than 400 features. In fact, good classification rates with about 50 features on average are achieved, with a 100% accuracy using the GA-LDA strategy for the GPDS database and 97% for the CASIA and IITD databases, approximately. For these last contact-less databases, reasonable EER rates are also obtained.

Probabilistic PCA Self-Organizing Maps

In this paper, we present a probabilistic neural model, which extends Kohonens self-organizing map (SOM) by performing a probabilistic principal component analysis (PPCA) at each neuron. Several SOMs have been proposed in the literature to capture the local principal subspaces, but our approach offers a probabilistic model while it has a low complexity on the dimensionality of the input space. This allows to process very high-dimensional data to obtain reliable estimations of the probability densities which are based on the PPCA framework. Experimental results are presented, which show the map formation capabilities of the proposal with high-dimensional data, and its potential in image and video compression applications.

Stochastic approximation for background modelling

Many background modelling approaches are based on mixtures of multivariate Gaussians with diagonal covariance matrices. This often yields good results, but complex backgrounds are not adequately captured, and post-processing techniques are needed. Here we propose the use of mixtures of uniform distributions and multivariate Gaussians with full covariance matrices. These mixtures are able to cope with both dynamic backgrounds and complex patterns of foreground objects. A learning algorithm is derived from the stochastic approximation framework, which has a very reduced computational complexity. Hence, it is suited for real time applications. Experimental results show that our approach outperforms the classic procedure in several benchmark videos.

Probabilistic Self-Organizing Maps for Continuous Data

The original self-organizing feature map did not define any probability distribution on the input space. However, the advantages of introducing probabilistic methodologies into self-organizing map models were soon evident. This has led to a wide range of proposals which reflect the current emergence of probabilistic approaches to computational intelligence. The underlying estimation theories behind them derive from two main lines of thought: the expectation maximization methodology and stochastic approximation methods. Here, we present a comprehensive view of the state of the art, with a unifying perspective of the involved theoretical frameworks. In particular, we examine the most commonly used continuous probability distributions, self-organization mechanisms, and learning schemes. Special emphasis is given to the connections among them and their relative advantages depending on the characteristics of the problem at hand. Furthermore, we evaluate their performance in two typical applications of self-organizing maps: classification and visualization.

Features for stochastic approximation based foreground detection

Abstract Foreground detection algorithms have sometimes relied on rather ad hoc procedures, even when probabilistic mixture models are defined. Moreover, the fact that the input features have different variances and that they are not independent from each other is often neglected, which hampers performance. Here we aim to obtain a background model which is not tied to any particular choice of features, and that accounts for the variability and the dependences among features. It is based on the stochastic approximation framework. A possible set of features is presented, and their suitability for this problem is assessed. Finally, the proposed procedure is compared with several state-of-the-art alternatives, with satisfactory results.

DYNAMIC COMPETITIVE PROBABILISTIC PRINCIPAL COMPONENTS ANALYSIS

We present a new neural model which extends the classical competitive learning (CL) by performing a Probabilistic Principal Components Analysis (PPCA) at each neuron. The model also has the ability to learn the number of basis vectors required to represent the principal directions of each cluster, so it overcomes a drawback of most local PCA models, where the dimensionality of a cluster must be fixed a priori. Experimental results are presented to show the performance of the network with multispectral image data.