Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2604.18940 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908982603415552 |
|---|---|
| author | Yong, Jiawei Qu, Deyuan Chen, Qi Oguchi, Kentaro Fukushima, Shintaro |
| author_facet | Yong, Jiawei Qu, Deyuan Chen, Qi Oguchi, Kentaro Fukushima, Shintaro |
| contents | Autonomous driving systems often degrade under adverse visibility conditions-such as rain, nighttime, or snow-where online scene geometry (e.g., lane dividers, road boundaries, and pedestrian crossings) becomes sparse or fragmented. While high-definition (HD) maps can provide missing structural context, they are costly to construct and maintain at scale. We propose Localization-Guided Foreground Augmentation (LG-FA), a lightweight and plug-and-play inference module that enhances foreground perception by enriching geometric context online. LG-FA: (i) incrementally constructs a sparse global vector layer from per-frame Bird's-Eye View (BEV) predictions; (ii) estimates ego pose via class-constrained geometric alignment, jointly improving localization and completing missing local topology; and (iii) reprojects the augmented foreground into a unified global frame to improve per-frame predictions. Experiments on challenging nuScenes sequences demonstrate that LG-FA improves the geometric completeness and temporal stability of BEV representations, reduces localization error, and produces globally consistent lane and topology reconstructions. The module can be seamlessly integrated into existing BEV-based perception systems without backbone modification. By providing a reliable geometric context prior, LG-FA enhances temporal consistency and supplies stable structural support for downstream modules such as tracking and decision-making. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_18940 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Localization-Guided Foreground Augmentation in Autonomous Driving Yong, Jiawei Qu, Deyuan Chen, Qi Oguchi, Kentaro Fukushima, Shintaro Computer Vision and Pattern Recognition Robotics Autonomous driving systems often degrade under adverse visibility conditions-such as rain, nighttime, or snow-where online scene geometry (e.g., lane dividers, road boundaries, and pedestrian crossings) becomes sparse or fragmented. While high-definition (HD) maps can provide missing structural context, they are costly to construct and maintain at scale. We propose Localization-Guided Foreground Augmentation (LG-FA), a lightweight and plug-and-play inference module that enhances foreground perception by enriching geometric context online. LG-FA: (i) incrementally constructs a sparse global vector layer from per-frame Bird's-Eye View (BEV) predictions; (ii) estimates ego pose via class-constrained geometric alignment, jointly improving localization and completing missing local topology; and (iii) reprojects the augmented foreground into a unified global frame to improve per-frame predictions. Experiments on challenging nuScenes sequences demonstrate that LG-FA improves the geometric completeness and temporal stability of BEV representations, reduces localization error, and produces globally consistent lane and topology reconstructions. The module can be seamlessly integrated into existing BEV-based perception systems without backbone modification. By providing a reliable geometric context prior, LG-FA enhances temporal consistency and supplies stable structural support for downstream modules such as tracking and decision-making. |
| title | Localization-Guided Foreground Augmentation in Autonomous Driving |
| topic | Computer Vision and Pattern Recognition Robotics |
| url | https://arxiv.org/abs/2604.18940 |