Explanation Retention Under Quantization and Temporal Reuse in Real-Time Object Detection: A YOLOv10x Study on Edge Platforms
Electronics · 20.11.2026 · 15 (22) · 5420
DOI: 10.3390/electronics15225420 ↗Kurz gefasst
Fachartikel in der Zeitschrift Electronics, online erschienen am 20.11.2026. Verfasst von Surbhi Bahri und 1 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Explanations released by a real-time object detector are often validated at full precision and assumed to remain stable after edge quantization. We test this assumption for a frozen YOLOv10x detector with a fused Grad-CAM teacher and a distilled forward-only surrogate under FP16 and INT8 deployment in five regimes from BDD100K and VisDrone-VID2019. Detection retention is measured by relative mAP and explanation retention by agreement between deployed and full-precision surrogate maps. Under INT8, explanation agreement fell by 5.8–11.4%, 5.4–7.3 times the noise floor of a 1/255 input perturbation, while mAP fell by 1.7–7.1%; on pairs not used to develop the surrogate, the loss was 9.6%. Controls attributed about three quarters of the loss to the feature-change step of a fixed-box-first decomposition, and showed that the asymmetry persisted on training-excluded sequences and across training seeds and calibration sets and that the surrogate’s maps depend on the learned weights. Two regimes repeated on a Jetson AGX Orin gave similar results. A motion-aware transport operator gave model-estimated reuse ages from 26 frames in daylight to six in dense traffic. Accuracy alone therefore does not establish that a detector’s explanations are retained after deployment.
Zitieren
APA
Bahri, S., & Mouftah, H. T. (2026). Explanation Retention Under Quantization and Temporal Reuse in Real-Time Object Detection: A YOLOv10x Study on Edge Platforms. Electronics, 15(22), 5420. https://doi.org/10.3390/electronics15225420
Vancouver
Bahri S, Mouftah HT. Explanation Retention Under Quantization and Temporal Reuse in Real-Time Object Detection: A YOLOv10x Study on Edge Platforms. Electronics. 2026;15(22):5420. doi:10.3390/electronics15225420
BibTeX
@article{Bahri2026,
author = {Bahri, Surbhi and Mouftah, Hussein T.},
title = {Explanation Retention Under Quantization and Temporal Reuse in Real-Time Object Detection: A YOLOv10x Study on Edge Platforms},
journal = {Electronics},
year = {2026},
volume = {15},
number = {22},
pages = {5420},
doi = {10.3390/electronics15225420}
}
EUVIMED