Juan Manuel García

Hand gesture recognition following the dynamics of a topology-preserving network

We present a new structure capable of characterizing hand posture, as well as its movement. Topology of a self-organizing neural network determines posture, whereas its adaptation dynamics throughout time determines gesture. This adaptive character of the network allows us to avoid the correspondence problem of other methods, so that the gestures are modelled by the movement of the neurons. To validate this method, we have trained the system with 12 gestures, some of which are very similar, and have obtained high success rates (over 97%). This application of a self-organizing network opens up a new field of research because its topology is used to characterize the objects and not to classify them, as is usually the case.

Improvement of the Discrete Cosine Transform calculation by means of a recursive method

This paper presents a recursive method for function evaluation. The proposal argues for the use of a more complete primitive, namely a weighted sum, which converts the calculation of the function values into a recursive operation defined by a two input table. The weighted sum can be tuned for different values of the weighting parameters that hold the features of the concrete evaluated function. The calculation of the Discrete Cosine Transform (DCT) is improved by this method which provides a decrease of the operation count as well as an acceptable error bound. These results have repercussions on the compression standard JPEG, for the purposes of storage and transmission in web-based applications

A Knowledge Based Framework to Support Active Aging at Home Based Environments

Information and Communication Technologies can support Active Aging strategies in a scenario like the Smart Home. This paper details a person centered distributed framework, called TALISMAN+, whose aim is to promote personal autonomy by taking advantage of knowledge based technologies, sensors networks, mobile devices and internet. The proposed solution can support an elderly person to keep living alone at his house without being obliged to move to a residential center. The framework is composed by five subsystems: a reasoning module that is able to take local decisions at home in order to support active aging, a biomedical variables telemonitorisation platform running on a mobile device, a hybrid reasoning middleware aimed to assess cardiovascular risk in a remote way, a private vision based sensor subsystem, and a secure telematics solution that guarantees confidentiality for personal information. TALISMAN+ framework deployment is being evaluated at a real environment like the Accessible Digital Home.

Growing neural gas for vision tasks with time restrictions

Self-organizing neural networks try to preserve the topology of an input space by means of their competitive learning. This capacity is being used for the representation of objects and their motion. In addition, these applications usually have real-time constraints. In this work, diverse variants of a self-organizing network, the Growing Neural Gas, that allow an acceleration of the learning process are considered. However, this increase of speed causes that, in some cases, topology preservation is lost and, therefore, the quality of the representation. So, we have made a study to quantify topology preservation using different measures to establish the most suitable learning parameters, depending on the size of the network and on the available time for its adaptation.

Computational framework for behavioural modelling of neural subsystems

This paper presents a new approach to the problem of modelling living system dynamics. Our point of view claims the fact that behind the biological apparent complexity, a hidden simplicity may appear when a suitable modelling is developed. The framework is inspired on the computing features of biological systems by involving a set of elementary standard behaviours that can be combined in order to emulate more complex behaviours. The algebraic formalization is based on both a recursive primitive operation defined by a table which models the elementary behaviours and a multilevel operating mode that carries out behaviour combinations. A parametric architecture implements the model, providing a good trade-off between time delay calculation and memory requirements. In this paper, the simulation of neural subsystems is considered as an application. The comparison with other simulation techniques outlines the capabilities of our method to provide an accurate modelling together with a very simple circuit implementation.

Characterization and synthesis of objects using growing neural gas

In this article it is made a study of the characterization capacity and synthesis of objects of the self-organizing neural models. These networks, by means of their competitive learning, try to preserve the topology of an input space. This capacity is being used for the representation of objects and their movement with topology preserving networks. We characterized the object to represent by means of the obtained maps and kept information solely on the coordinates and the colour from the neurons. From this information it is made the synthesis of the original images, applying mathematical morphology and simple filters on the information which it is had.

Morphological applications for maps construction in path planning tasks

The process of taking decisions in an autonomous system has a strong dependence on the information received from the environment in which it works. The path planning process for a mobile robot is a subset of the problems that have to be solved in an automatic system of decisions, where the appropriateness of the method is roughly constrained not only by the achievement of a well-suited information from the environment, but also by how this information is acquired. The main objective of this work is to describe a robust model for robotic path planning in unknown environments. The proposed general model is based on obtaining secure paths where the robot will be able to move so as to complete its task. Thus, from an unknown location, the required information is collected by employing a mobile robot that moves freely through a real world with some generic behaviors that let the robot explore the world.

A computational framework based on behavioural modelling: Application to the matching of electrocardiogram (ECG) recordings

We present a computational framework to aid in cardiovascular disease diagnosis. Our method defines a set of standard behaviours obtained by the recursive calculation of a parameterized formula, and these behaviours are used to match the electrocardiogram (ECG) recording. The advantage of this proposal is the capability to extract from the huge set of numerical values of the ECGs a characteristic reduced pattern with behavioural meaning which allows more accurate and easy matching with unknown ECGs in order to diagnose the patients.

A morphological proposal for vision-based path planning

Many different path planning methods have been proposed over recent years, although there are only a few that deal with computer vision techniques. In this work we implement a path planning algorithm which takes into account a vision processing system. Thus, we develop a method that uses Mathematical Morphology to provide near-optimal paths throughout an environment. The experiments show that our path planning algorithm is able to locate good solution paths after a training process, which is necessary to fix some parameters. This will make possible its adaptation to a practical robot system.This work has been supported by the Spanish MCYT, project DPI200204434C0401 and by the Generalitat Valenciana, projects GV04B685,GV04B634.