Sistema de anotación semiautomático de señales electrocardiográficas de esfuerzo

Abstract

Despite the availability of electrocardiographic (ECG) databases does the medical sector not yet possesses an annotated stress database. This paper introduces an annotation system using RR-time series obtained from an eight lead stress database (DICARDIA). First, the system proposes to the user a lead (reference channel) according to its statistical measures. The user then undergoes a visual inspection validating or denying the channel proposed by the system. Afterwards, the system proposes two options based on the quality of the RR-time series. If the series contains few artifacts the annotations are realized using the interval of annotations; in the case of noisy series the system allows annotations beat by beat. The preliminary results realized over 172025 beats (approximately 15% of DICARDIA database) give a sensibility and positive predictive value of 97.66% and 96.71% respectively. The system will permit the delineation of stress databases, which will be an important starting point for evaluating the performance of QRS detectors.