Embedded implementation of biomedical applications in heterogeneous systems

Article type: Research Article

Authors: Latif, Rachid | Saddik, Amine^{; *}

Affiliations: Laboratory of Systems Engineering and Information Technology, ENSA, Ibn Zohr University, Agadir, Morocco

Correspondence: [*] Corresponding author. E-mail: A.saddik.ma@gmail.com.

Abstract: The high flexibility of the heterogeneous system based on CPU-GPU and CPU-FPGAoffer us huge performance in terms of processing time and resource optimization. We can use these systems in all areas provided that we have studied the adequacy between the algorithm and the architecture, this adequacy technique is called hardware software co-design. Today, the need for embedded systems is an irreplaceable reality, especially when we talk about real-time systems. Among the most well-known embedded systems are FPGAs, this system is based on the parallelization of algorithms via the pipline principle. In this work, we present the application of the heterogeneous system in the biomedical field, in particular the cardiac signals provided by the electrocardiogram and the different ECG signal filtering techniques. And also, we will proposing an embedded implementation of the ADTF filter in a homogeneous architecture (FPGA). And also, the heterogeneous architecture via the HLS OpenCL.

Keywords: Heterogeneous system, CPU-GPU, CPU-FPGA, ECG signal

DOI: 10.3233/BSI-200192

Journal: Biomedical Spectroscopy and Imaging, vol. 8, no. 3-4, pp. 73-80, 2019