Multi-channel heart-beat detection

In this paper, we propose a heart-rate measurement system employing triple sensing mechanisms (face color changes, finger color changes, and heart sound measurment) on an iPhone. The three proposed measurement systems each provide an independent heart-rate estimate as well as a combined heart-rate estimate based on the fusion of the individual sensors. Obtaining a single representation of the cardiac cycle from these three modalities requires data-level fusion. In the proposed method, we introduce a new heart-rate measurement enhancement method based on the Empirical Mode Decomposition (EMD). To measure the performance of the proposed method, we collected data on 56 subjects. Our experimental results show that the Root Mean Square Error (as compared to the result from a pulse oximeter) from our proposed method is 3.6 which provides better performance than traditional filtering in terms of noise reduction. Furthermore, this fused result is more accurate than any single modality.