FPG-NAS: FLOPs-Aware Gated Differentiable Neural Architecture Search for Efficient 6DoF Pose Estimation

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Ousalah, Nassim Ali, Rostami, Peyman, Kacem, Anis, Ghorbel, Enjie, Koumandakis, Emmanuel, Aouada, Djamila
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866917210709032960
author Ousalah, Nassim Ali
Rostami, Peyman
Kacem, Anis
Ghorbel, Enjie
Koumandakis, Emmanuel
Aouada, Djamila
author_facet Ousalah, Nassim Ali
Rostami, Peyman
Kacem, Anis
Ghorbel, Enjie
Koumandakis, Emmanuel
Aouada, Djamila
contents We introduce FPG-NAS, a FLOPs-aware Gated Differentiable Neural Architecture Search framework for efficient 6DoF object pose estimation. Estimating 3D rotation and translation from a single image has been widely investigated yet remains computationally demanding, limiting applicability in resource-constrained scenarios. FPG-NAS addresses this by proposing a specialized differentiable NAS approach for 6DoF pose estimation, featuring a task-specific search space and a differentiable gating mechanism that enables discrete multi-candidate operator selection, thus improving architectural diversity. Additionally, a FLOPs regularization term ensures a balanced trade-off between accuracy and efficiency. The framework explores a vast search space of approximately 10\textsuperscript{92} possible architectures. Experiments on the LINEMOD and SPEED+ datasets demonstrate that FPG-NAS-derived models outperform previous methods under strict FLOPs constraints. To the best of our knowledge, FPG-NAS is the first differentiable NAS framework specifically designed for 6DoF object pose estimation.
format Preprint
id arxiv_https___arxiv_org_abs_2508_03618
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle FPG-NAS: FLOPs-Aware Gated Differentiable Neural Architecture Search for Efficient 6DoF Pose Estimation
Ousalah, Nassim Ali
Rostami, Peyman
Kacem, Anis
Ghorbel, Enjie
Koumandakis, Emmanuel
Aouada, Djamila
Computer Vision and Pattern Recognition
We introduce FPG-NAS, a FLOPs-aware Gated Differentiable Neural Architecture Search framework for efficient 6DoF object pose estimation. Estimating 3D rotation and translation from a single image has been widely investigated yet remains computationally demanding, limiting applicability in resource-constrained scenarios. FPG-NAS addresses this by proposing a specialized differentiable NAS approach for 6DoF pose estimation, featuring a task-specific search space and a differentiable gating mechanism that enables discrete multi-candidate operator selection, thus improving architectural diversity. Additionally, a FLOPs regularization term ensures a balanced trade-off between accuracy and efficiency. The framework explores a vast search space of approximately 10\textsuperscript{92} possible architectures. Experiments on the LINEMOD and SPEED+ datasets demonstrate that FPG-NAS-derived models outperform previous methods under strict FLOPs constraints. To the best of our knowledge, FPG-NAS is the first differentiable NAS framework specifically designed for 6DoF object pose estimation.
title FPG-NAS: FLOPs-Aware Gated Differentiable Neural Architecture Search for Efficient 6DoF Pose Estimation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2508.03618