Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2504.18112 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909593954680832 |
|---|---|
| author | Ghimire, Deepak Kim, Byoungjun Kim, Donghoon Jeong, SungHwan |
| author_facet | Ghimire, Deepak Kim, Byoungjun Kim, Donghoon Jeong, SungHwan |
| contents | Road surface reconstruction plays a crucial role in autonomous driving, providing essential information for safe and smooth navigation. This paper enhances the RoadBEV [1] framework for real-time inference on edge devices by optimizing both efficiency and accuracy. To achieve this, we proposed to apply Isomorphic Global Structured Pruning to the stereo feature extraction backbone, reducing network complexity while maintaining performance. Additionally, the head network is redesigned with an optimized hourglass structure, dynamic attention heads, reduced feature channels, mixed precision inference, and efficient probability volume computation. Our approach improves inference speed while achieving lower reconstruction error, making it well-suited for real-time road surface reconstruction in autonomous driving. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2504_18112 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Study on Real-Time Road Surface Reconstruction Using Stereo Vision Ghimire, Deepak Kim, Byoungjun Kim, Donghoon Jeong, SungHwan Computer Vision and Pattern Recognition Road surface reconstruction plays a crucial role in autonomous driving, providing essential information for safe and smooth navigation. This paper enhances the RoadBEV [1] framework for real-time inference on edge devices by optimizing both efficiency and accuracy. To achieve this, we proposed to apply Isomorphic Global Structured Pruning to the stereo feature extraction backbone, reducing network complexity while maintaining performance. Additionally, the head network is redesigned with an optimized hourglass structure, dynamic attention heads, reduced feature channels, mixed precision inference, and efficient probability volume computation. Our approach improves inference speed while achieving lower reconstruction error, making it well-suited for real-time road surface reconstruction in autonomous driving. |
| title | Study on Real-Time Road Surface Reconstruction Using Stereo Vision |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2504.18112 |