**5. Conclusions**

In this paper, we proposed an end-to-end deep learning approach fusing fingerprint and ECG signals for boosting the PAD capabilities of fingerprint biometrics. We also introduced EfficientNet, a state-of-the-art network, for learning efficient fingerprint feature representations. The experimental results prove the superiority of the EfficientNet over other known pre-trained CNNs in terms of accuracy and efficiency. For the ECG signals, we proposed three different architectures and configurations FC, a 1D-CNN, and a 2D-CNN. With the 2D-CNN model, we transformed the ECG features into images using a generator network. The experimental results obtained on a multimodal dataset composed of fingerprint and ECG signals reveal the promising capability of the proposed solution in terms of the classification accuracy and computation time. Although we used a customized database of fingerprint and ECG signal to validate the proposed method, we intend to use a database of fingerprint and ECG signal captured simultaneously using a multimodal sensor. Since this type of sensor is not commercially available, we would like to develop a prototype of such a sensor, create a database of real multimodal data, and use it for the validation of the proposed method as our future work.

**Author Contributions:** Data curation, M.S.I.; methodology, R.M.J. and Y.B.; supervision, H.M.; writing—original draft, R.M.J.; writing—review and editing, H.M., Y.B., and M.S.I. All authors have read and agreed to the published version of the manuscript.

**Funding:** The authors would like to thank Deanship of scientific research for funding and supporting this research through the initiative of DSR Graduate Students Research Support (GSR).

**Acknowledgments:** The authors thank the Deanship of Scientific Research and RSSU at King Saud University for their technical support.

**Conflicts of Interest:** The authors declare no conflicts of interest.
