Methodology for the automatic interpretation of electrocardiograms, based on the trigonometric discrimination of maximal modulus of the wave transformation, using neural networks

Jorge Alejandro CASTILLO MORALES student laureate
jcastillo@upb.edu

°1976 Bolivia System engineering, Universidad Privada Boliviana, Cochabamba, Bolivia, 2001

The development of new technology should not be the sole prerogative of the Western nations. Castillo Morales’ thesis demonstrates that very sound and advanced research into new technology is possible in developing nations. This is a very clearly written and well planned work that makes use of modern technology. This study uses complex computer technology (neural networks) for the automatic interpretation of electrocardiograms. The electrocardiogram (ECG) is an important form of technical monitoring that is used to evaluate patients with cardiac conditions. However, the interpretation of an ECG requires a considerable amount of expertise, and needs to be performed by doctors with the requisite experience and training. Should such doctors not be available, automatic analysis systems represent a good alternative. The author researched the reliability of pattern recognition in the ECG by neural networks. This proves to be very successful, if the systems he devised are applied. On the basis of these data, a prototype device has already been constructed which can be tested for clinical applications. The advantage of the system devised by the author is that it turns out to be cheaper than existing automatic analysis systems. This could make the application of this technology easier in developing countries, and more accessible in more remote areas. In this way, such systems could contribute to an improvement in the quality of health care in developing countries.

report by Prof. L. Mertens, Congenital cardiology, University Hospital Leuven, Belgium