This paper describes the development of fully connected artificial neural network (RSL_ANN), receiving 19 ECG features (11 morphological, 4 on F waves and 4 on heart-rate variability). The network was created and tested on 8028 annotated ECGs acquired with the Kardia device. Less than 3% of the ECGs included in the database could not be used in this study due to high levels of noise. Performance of RSL_ANN was very good and very similar in all datasets, with AUC over 90%. The work shows the value of Kardia for providing high volumes of quality data to aid the development of advanced diagnostic algorithms for AF.