EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC-ND 4.0

Exploring FINN Neural Network Accelerators with ARM Hardcore and RISC-V Softcore Processors in SRAM-based FPGA

Arthur F. Ely, Vicente Veiga, Fabio Benevenuti, Antônio Carlos Beck, José Rodrigo Azambuja, Fernanda Kastensmidt

Journal of Integrated Circuits and Systems · 12.04.2026 · 21 (1) · 1-14

DOI: 10.29292/jics.v21i1.1028 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Integrated Circuits and Systems, online erschienen am 12.04.2026. Verfasst von Arthur F. Ely und 5 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

This study presents a detailed exploration of Convolutional Neural Networks (CNNs) for aerial image classification on satellite platforms, leveraging the FINN framework implemented on AMD/Xilinx 7-series APSoCs. By systematically examining the trade-offs between quantization levels, network topologies, and architectural parallelism, we achieved high accuracy for different quantizations, and demonstrated robust performance scaling from 542 FPS to 16.3k FPS. Our results highlight critical design decisions for integrating CNN accelerators with both hardcore ARM Cortex-A9 and softcore NOEL-V RISC-V processors, offering insights into resource utilization, energy efficiency, and system performance. This work provides a valuable roadmap for optimizing in-orbit image processing systems in next-generation satellite applications.

Zitieren

APA

Ely, A. F., Veiga, V., Benevenuti, F., Beck, A. C., Azambuja, J. R., & Kastensmidt, F. (2026). Exploring FINN Neural Network Accelerators with ARM Hardcore and RISC-V Softcore Processors in SRAM-based FPGA. Journal of Integrated Circuits and Systems, 21(1), 1-14. https://doi.org/10.29292/jics.v21i1.1028

Vancouver

Ely AF, Veiga V, Benevenuti F, Beck AC, Azambuja JR, Kastensmidt F. Exploring FINN Neural Network Accelerators with ARM Hardcore and RISC-V Softcore Processors in SRAM-based FPGA. Journal of Integrated Circuits and Systems. 2026;21(1):1-14. doi:10.29292/jics.v21i1.1028

BibTeX
@article{Ely2026,
  author = {Ely, Arthur F. and Veiga, Vicente and Benevenuti, Fabio and Beck, Antônio Carlos and Azambuja, José Rodrigo and Kastensmidt, Fernanda},
  title = {Exploring FINN Neural Network Accelerators with ARM Hardcore and RISC-V Softcore Processors in SRAM-based FPGA},
  journal = {Journal of Integrated Circuits and Systems},
  year = {2026},
  volume = {21},
  number = {1},
  pages = {1--14},
  doi = {10.29292/jics.v21i1.1028}
}
Kostenlos starten →