Uncertainty-Instructed Structure Injection for Generalizable HD Map Construction

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Liu, Xiaolu, Yang, Ruizi, Wang, Song, Li, Wentong, Chen, Junbo, Zhu, Jianke
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866908289858535424
author Liu, Xiaolu
Yang, Ruizi
Wang, Song
Li, Wentong
Chen, Junbo
Zhu, Jianke
author_facet Liu, Xiaolu
Yang, Ruizi
Wang, Song
Li, Wentong
Chen, Junbo
Zhu, Jianke
contents Reliable high-definition (HD) map construction is crucial for the driving safety of autonomous vehicles. Although recent studies demonstrate improved performance, their generalization capability across unfamiliar driving scenes remains unexplored. To tackle this issue, we propose UIGenMap, an uncertainty-instructed structure injection approach for generalizable HD map vectorization, which concerns the uncertainty resampling in statistical distribution and employs explicit instance features to reduce excessive reliance on training data. Specifically, we introduce the perspective-view (PV) detection branch to obtain explicit structural features, in which the uncertainty-aware decoder is designed to dynamically sample probability distributions considering the difference in scenes. With probabilistic embedding and selection, UI2DPrompt is proposed to construct PV-learnable prompts. These PV prompts are integrated into the map decoder by designed hybrid injection to compensate for neglected instance structures. To ensure real-time inference, a lightweight Mimic Query Distillation is designed to learn from PV prompts, which can serve as an efficient alternative to the flow of PV branches. Extensive experiments on challenging geographically disjoint (geo-based) data splits demonstrate that our UIGenMap achieves superior performance, with +5.7 mAP improvement on the nuScenes dataset. Source code will be available at https://github.com/xiaolul2/UIGenMap.
format Preprint
id arxiv_https___arxiv_org_abs_2503_23109
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Uncertainty-Instructed Structure Injection for Generalizable HD Map Construction
Liu, Xiaolu
Yang, Ruizi
Wang, Song
Li, Wentong
Chen, Junbo
Zhu, Jianke
Computer Vision and Pattern Recognition
Reliable high-definition (HD) map construction is crucial for the driving safety of autonomous vehicles. Although recent studies demonstrate improved performance, their generalization capability across unfamiliar driving scenes remains unexplored. To tackle this issue, we propose UIGenMap, an uncertainty-instructed structure injection approach for generalizable HD map vectorization, which concerns the uncertainty resampling in statistical distribution and employs explicit instance features to reduce excessive reliance on training data. Specifically, we introduce the perspective-view (PV) detection branch to obtain explicit structural features, in which the uncertainty-aware decoder is designed to dynamically sample probability distributions considering the difference in scenes. With probabilistic embedding and selection, UI2DPrompt is proposed to construct PV-learnable prompts. These PV prompts are integrated into the map decoder by designed hybrid injection to compensate for neglected instance structures. To ensure real-time inference, a lightweight Mimic Query Distillation is designed to learn from PV prompts, which can serve as an efficient alternative to the flow of PV branches. Extensive experiments on challenging geographically disjoint (geo-based) data splits demonstrate that our UIGenMap achieves superior performance, with +5.7 mAP improvement on the nuScenes dataset. Source code will be available at https://github.com/xiaolul2/UIGenMap.
title Uncertainty-Instructed Structure Injection for Generalizable HD Map Construction
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2503.23109