E. Ros Vidal - Winner of the Award at Como Workshop on Biosignal Interpertation

Eduardo Ros Vidal received his B.Sc. Degree in Physics in 1993, Electronic Engineering in 1996, and the Ph.D. degree in 1997, all from the University of Granada, Spain. After receiving his PhD, he spent 4 months at the Department of Electronic Engineering, King's College of London (UK). Currently, he is associate professor at the Departamento de Arquitectura y Tecnología de Computadores. He teaches at the Informatics Engineering School, mainly the subject "Introduction to Computers", and the PhD. courses called "Biomedical Applications" and "Computers and Biomedical Instrumentation".

Eduardo's PhD Thesis is entitled "VLSI implementation of Neural Structures inspired in the biology". It addressed some topics related to the research field known as Neuromorphic Engineering. Biological neurons were studied in order to discriminate the properties that lead to their computational power within the nervous systems of other properties that seemed not to be significant for computational tasks. After this functional abstraction, the research work was concentrated in evaluating whether bio-inspired circuits, built with standard VLSI technology, can take advantage of similar computational primitives for certain processing tasks [1-8].

Currently he and his group continue this line of research with three projects in which he directly collaborates:

ECOVISION: Artificial Vision System Based on Early Cognitive Cortical Processing
SPIKEFORCE: Real-time Spiking Networks for Robot Control
CORTIVIS: Cortical Visual Neuroprosthesis for the Blind

In ECOVISION [9] the research group attempt to build (Hardware/Software) an artificial vision system inspired in the higher vertebrates visual systems that exhibits processing performance levels still far ahead the ones achieved by artificial machines. CORTIVIS [10] addresses the implementation of a neuroprosthesis for the blind. Using conventional CCDs or more bio-inspired front-ends they try to emulate the whole visual pathway, preserving the information integrity, in order to transfer it directly to the visual cortex through a micro-array of neuro-stimulators. Finally, SPIKEFORCE [11] is focussed on the study of the cerebellum, our role in this project is to evaluate the possibility of using the same computational primitives (based on spiking neurons) observed in the cerebellum to gain efficiency and flexibility in robot control applications. In all these applications they design digital systems (FPGAs) to achieve real-time processing.

More directly related with signal processing for biomedical applications, the group have studied the problem of Paroxysmal Atrial Fibrillation Diagnosis and On-set episode prediction [12, 13] based on ECG traces not explicitly containing fibrillation episodes. This challenging problem was proposed as an international initiative [14] to encourage different research groups to participate in this interesting field. The research group has also addressed this application as a Multi-objective optimisation problem using Genetic Algorithms [15].

Main publications:

[1]Pelayo, F. J.; Ros, E.; Martín-Smith, P.; Fernández, F.J.; Prieto, A.: A VLSI approach to the implementation of Additive and Shunting Neural Networks, "Lecture Notes in Computer Science 930", Springer Verlag, pp. 728-735, 1995.
[2]Anguita, M.; Pelayo, F.J.; Ros, E.; Palomar, D.; Prieto A.: VLSI Implementations of CNNs for Image Processing and Vision Tasks: Single and Multiple Chip Approaches, "Proceedings of the Fourth IEEE International Workshop on Cellular Neural Networks and their Applications", Sevilla, pp. 479-484, Jun. 24-26 de 1996.
[3]Pelayo, F.J.; Ros, E.; Arreguit, X.; Prieto, A.: VLSI Implementation of a Neural Model Using Spikes, "Analog Integrated Circuits and Signal Processing", Special Issue on Neuromorphic Engineering, Kluwer Academic Publishers, Vol. 13, pp. 111-121, 1997.
[4]Ros, E., Pelayo, F.J., Pino, B., and Prieto, B.: Firing Rate and Phase Coding Circuits for Neural Computation using Spikes, MicroNeuro'97, Microelectronics for Neural Networks, Evolutionary & Fuzzy Systems, pp. 305-311, September, 1997.
[5]Anguita, M.; Pelayo, F.J.; Ros, E.; Palomar, D.; Prieto, A.: Single and Multiple Chip Approaches to Cellular Structures for Vision Tasks and Image Processing, "Analog Intagrated Circuits and Signal Processing", Kluwer Academic Publishers, Vol. 15, pp. 263-275, 1998.
[6]Martin-Smith, P.; Pelayo, F.J.; Ros, E; Prieto, A: Supervised VQ Learning based on Temporal Inhibition, Foundations and Tools for Neural Modeling, Lecture Notes in Computer Science, Springer Verlag, Vol. 1606, pp. 610-620, 1999.
[7]Ros, E.; Pelayo, F.J.;Rojas, I.; Fernández, F.J.; Prieto, A.: A VLSI Approach for Spike Timing Coding, Engineering Applications of Bio-Inspired Artificial Neural Networks, Lecture Notes in Computer Science, Springer Verlag, Vol. 1607, pp. 166-175, 1999.
[8]Ros, E.; Pelayo, F.J.; Martín-Smith, P.; Palomar, D.; Prieto, A.: Competitive and Temporal Inhibition Structures with Spiking Neurons, Neural Processing Letters, Kluwer Academic Publishers, Vol. 11, 197-208, 2000.
[9] www.pspc.dibe.unige.it/ecovision/
[10] http://cortivis.umh.es/
[11] www.spikeforce.org/
[12]Mota, S.; Ros, E.; Fernández, F.J.; Díaz, A.F.; Prieto, A.: ECG Parameter Characterization of Paroxysmal Atrial Fibrillation, pp. 247-250, IV International Workshop on Biosignal Interpretation (BSI 2002), Como, June 2002.
[13]Ros, E.; Mota, S.; Toro, F.J.; Díaz, A.F.; Fernández, F.J.: Paroxysmal Atrial Fibrillation: Automatic Diagnosis Algorithm Based on not Fibrillating ECGs, pp. 251-254, IV International Workshop on Biosignal Interpretation (BSI 2002), Como, June 2002.
[14] www.cinc.org/LocalHost/CIC2001_1.htm
[15]Toro, F.J.; Ros, E.; Mota, S.; Ortega, J.; Prieto, A.: Multi-objective Optimization for Paroxysmal Atrial Fibrillation Diagnosis, Accepted for the IBERMAIA'2002 Sevilla.

For further information visit Eduardo Ros's web-site: http://atc.ugr.es/~eduardo/