Saved in:
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2411.16072 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911082817257472 |
|---|---|
| author | Yu, Zhu Pang, Bowen Liu, Lizhe Zhang, Runmin Li, Qiang Cao, Si-Yuan Luo, Maochun Chen, Mingxia Yang, Sheng Shen, Hui-Liang |
| author_facet | Yu, Zhu Pang, Bowen Liu, Lizhe Zhang, Runmin Li, Qiang Cao, Si-Yuan Luo, Maochun Chen, Mingxia Yang, Sheng Shen, Hui-Liang |
| contents | We introduce LOcc, an effective and generalizable framework for open-vocabulary occupancy (OVO) prediction. Previous approaches typically supervise the networks through coarse voxel-to-text correspondences via image features as intermediates or noisy and sparse correspondences from voxel-based model-view projections. To alleviate the inaccurate supervision, we propose a semantic transitive labeling pipeline to generate dense and fine-grained 3D language occupancy ground truth. Our pipeline presents a feasible way to dig into the valuable semantic information of images, transferring text labels from images to LiDAR point clouds and ultimately to voxels, to establish precise voxel-to-text correspondences. By replacing the original prediction head of supervised occupancy models with a geometry head for binary occupancy states and a language head for language features, LOcc effectively uses the generated language ground truth to guide the learning of 3D language volume. Through extensive experiments, we demonstrate that our transitive semantic labeling pipeline can produce more accurate pseudo-labeled ground truth, diminishing labor-intensive human annotations. Additionally, we validate LOcc across various architectures, where all models consistently outperform state-of-the-art zero-shot occupancy prediction approaches on the Occ3D-nuScenes dataset. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2411_16072 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Language Driven Occupancy Prediction Yu, Zhu Pang, Bowen Liu, Lizhe Zhang, Runmin Li, Qiang Cao, Si-Yuan Luo, Maochun Chen, Mingxia Yang, Sheng Shen, Hui-Liang Computer Vision and Pattern Recognition We introduce LOcc, an effective and generalizable framework for open-vocabulary occupancy (OVO) prediction. Previous approaches typically supervise the networks through coarse voxel-to-text correspondences via image features as intermediates or noisy and sparse correspondences from voxel-based model-view projections. To alleviate the inaccurate supervision, we propose a semantic transitive labeling pipeline to generate dense and fine-grained 3D language occupancy ground truth. Our pipeline presents a feasible way to dig into the valuable semantic information of images, transferring text labels from images to LiDAR point clouds and ultimately to voxels, to establish precise voxel-to-text correspondences. By replacing the original prediction head of supervised occupancy models with a geometry head for binary occupancy states and a language head for language features, LOcc effectively uses the generated language ground truth to guide the learning of 3D language volume. Through extensive experiments, we demonstrate that our transitive semantic labeling pipeline can produce more accurate pseudo-labeled ground truth, diminishing labor-intensive human annotations. Additionally, we validate LOcc across various architectures, where all models consistently outperform state-of-the-art zero-shot occupancy prediction approaches on the Occ3D-nuScenes dataset. |
| title | Language Driven Occupancy Prediction |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2411.16072 |