EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

A Hardware-Aware Analysis of PTQ and QAT Quantized CNNs for Object Detection on FPGA

Noura Jariri, Kaoutar Allabouche, Mohamed Benaly, Mohammed Chaman, Rania Majdoubi, Abdelkader Hadjoudja

Engineering, Technology & Applied Science Research · 04.04.2026 · 16 (2) · 34498-34504

DOI: 10.48084/etasr.17475 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Engineering, Technology & Applied Science Research, online erschienen am 04.04.2026. Verfasst von Noura Jariri und 5 weiteren Autor:innen. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 29 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Real-time object detection on embedded platforms is critical for safety-critical and industrial applications, but FPGA deployment remains challenging due to constraints on numerical precision, latency, and hardware resources. Although quantization is widely used to enable efficient FPGA inference, its impact on object-detection models combining classification and bounding-box regression has not been systematically analyzed within an hls4ml based workflow. This work compares Post-Training Quantization (PTQ) and Quantization-Aware Training (QAT) for deploying a lightweight CNN-based detector on an FPGA. An FP32 model is quantized to INT16 and INT8 using QKeras and subsequently converted to fixed-point hardware representations with hls4ml. The results show that PTQ severely degrades detection performance, reducing classification accuracy to approximately 10% and mean IoU below 0.30. In contrast, QAT preserves near-floating-point performance, achieving ≈94% accuracy and ≈0.89 IoU at the software level for both INT16 and INT8. However, default HLS fixed-point configurations introduce software-hardware discrepancies, particularly in classification. A regression-aware refinement that increases fractional precision in the bounding-box head restores hardware-level localization accuracy (IoU ≈0.89), while residual classification gaps remain due to fixed-point constraints. These findings demonstrate that reliable FPGA-based object detection requires both QAT and hardware-aware fixed-point design, providing practical guidelines for low-precision deployment using hls4ml.

Zitieren

APA

Jariri, N., Allabouche, K., Benaly, M., Chaman, M., Majdoubi, R., & Hadjoudja, A. (2026). A Hardware-Aware Analysis of PTQ and QAT Quantized CNNs for Object Detection on FPGA. Engineering, Technology & Applied Science Research, 16(2), 34498-34504. https://doi.org/10.48084/etasr.17475

Vancouver

Jariri N, Allabouche K, Benaly M, Chaman M, Majdoubi R, Hadjoudja A. A Hardware-Aware Analysis of PTQ and QAT Quantized CNNs for Object Detection on FPGA. Engineering, Technology & Applied Science Research. 2026;16(2):34498-34504. doi:10.48084/etasr.17475

BibTeX
@article{Jariri2026,
  author = {Jariri, Noura and Allabouche, Kaoutar and Benaly, Mohamed and Chaman, Mohammed and Majdoubi, Rania and Hadjoudja, Abdelkader},
  title = {A Hardware-Aware Analysis of PTQ and QAT Quantized CNNs for Object Detection on FPGA},
  journal = {Engineering, Technology & Applied Science Research},
  year = {2026},
  volume = {16},
  number = {2},
  pages = {34498--34504},
  doi = {10.48084/etasr.17475}
}
Kostenlos starten →