The results demonstrated that the proposed implementation
works well on the MIT-BIH databases. Especially the
high sensitivity and the low number of false negatives evidenced
that the approach was applicable, while the number
of false positives could still be reduced.