LA-Pose: Latent Action Pretraining Meets Pose Estimation

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Wang, Zhengqing, Nair, Saurabh, Chidananda, Prajwal, Kachana, Pujith, Li, Samuel, Brown, Matthew, Furukawa, Yasutaka
Natura: Preprint
Pubblicazione: 2026
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866914520086085632
author Wang, Zhengqing
Nair, Saurabh
Chidananda, Prajwal
Kachana, Pujith
Li, Samuel
Brown, Matthew
Furukawa, Yasutaka
author_facet Wang, Zhengqing
Nair, Saurabh
Chidananda, Prajwal
Kachana, Pujith
Li, Samuel
Brown, Matthew
Furukawa, Yasutaka
contents This paper revisits camera pose estimation through the lens of self-supervised pretraining, focusing on inverse-dynamics pretraining as a scalable alternative to the current trend of fully supervised training with 3D annotations. Concretely, we employ inverse- and forward-dynamics models to learn latent action representations, similar to Genie from large-scale driving videos. Our idea is simple yet effective. Existing methods use latent actions in their original capacity, that is, as action conditioning of world-models or as proxies of robot action parameters in policy networks. Our method, dubbed LA-Pose, repurposes the latent action features as inputs to a camera pose estimator, finetuned on a limited set of high-quality 3D annotations. This formulation enables accurate and generalizable pose prediction while maintaining feed-forward efficiency. Extensive experiments on driving benchmarks show that LA-Pose achieves competitive and even superior performance to state-of-the-art methods while using orders of magnitude less labeled data. Concretely, on the Waymo and PandaSet benchmarks, LA-Pose achieves over 10% higher pose accuracy than recent feed-forward methods. To our knowledge, this work is the first to demonstrate the power of inverse-dynamics self-supervised learning for pose estimation.
format Preprint
id arxiv_https___arxiv_org_abs_2604_27448
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle LA-Pose: Latent Action Pretraining Meets Pose Estimation
Wang, Zhengqing
Nair, Saurabh
Chidananda, Prajwal
Kachana, Pujith
Li, Samuel
Brown, Matthew
Furukawa, Yasutaka
Computer Vision and Pattern Recognition
This paper revisits camera pose estimation through the lens of self-supervised pretraining, focusing on inverse-dynamics pretraining as a scalable alternative to the current trend of fully supervised training with 3D annotations. Concretely, we employ inverse- and forward-dynamics models to learn latent action representations, similar to Genie from large-scale driving videos. Our idea is simple yet effective. Existing methods use latent actions in their original capacity, that is, as action conditioning of world-models or as proxies of robot action parameters in policy networks. Our method, dubbed LA-Pose, repurposes the latent action features as inputs to a camera pose estimator, finetuned on a limited set of high-quality 3D annotations. This formulation enables accurate and generalizable pose prediction while maintaining feed-forward efficiency. Extensive experiments on driving benchmarks show that LA-Pose achieves competitive and even superior performance to state-of-the-art methods while using orders of magnitude less labeled data. Concretely, on the Waymo and PandaSet benchmarks, LA-Pose achieves over 10% higher pose accuracy than recent feed-forward methods. To our knowledge, this work is the first to demonstrate the power of inverse-dynamics self-supervised learning for pose estimation.
title LA-Pose: Latent Action Pretraining Meets Pose Estimation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2604.27448