Research map / Detection and segmentation / 3D detection and driving
OpenSight: A Simple Open-Vocabulary Framework for LiDAR-Based Object Detection
Hu Zhang, xu jianhua, Tao Tang, Haiyang Sun, Xin Yu, Zi Helen Huang, Kaicheng Yu
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add temporal awareness to generic object perception (ablation) | OpenSight · nuScenes mAP-far | mAP-far: 22.1 -> 23.0 | Improved |
| add spatial awareness (object bank with LLM size priors) (ablation) | OpenSight · nuScenes mAP-far | mAP-far: 22.1 -> 22.8 | Improved |
| combine temporal and spatial awareness (ablation) | OpenSight · nuScenes mAP-far | mAP-far: 22.1 -> 23.5 | Improved |
| add 3D-2D feature integration (ablation) | OpenSight · nuScenes mAP-large | mAP-large: 9.4 -> 10.9 | Improved |
| add LLM size-prior post-processing (ablation) | OpenSight · nuScenes mAP-large | mAP-large: 10.9 -> 11.5 | Improved |
| use object-level Grounding DINO instead of image-level CLIP for semantics (ablation) | OpenSight · nuScenes mAP-near | mAP-near: 19.0 -> 28.0 | Improved |
| use known-class co-embedding paradigm (CoDA) instead of generic-object perception | open-vocabulary LiDAR detection · nuScenes novel classes | mAP: 23.5 -> 10.3 | Hurt |
| OpenSight on large objects compared with CoDA | OpenSight · nuScenes mAP-large | mAP-large: 17.9 -> 11.5 | Hurt |
| OpenSight on trailer class | OpenSight · nuScenes trailer | AP: 0.8 vs box-only supervised 0.7 | No clear effect |
Later papers that compared with this method
1 later accepted paper reports OpenSight as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- OV-SCAN: Semantically Consistent Alignment for Novel Object Discovery in Open-Vocabulary 3D Object Detection (ICCV 2025 · better on 2 of 2 shared results)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Find n' Propagate: Open-Vocabulary 3D Object Detection in Urban Environments (ECCV 2024 · similarity 0.91)
- Open-CRB: Toward Open World Active Learning for 3D Object Detection (TPAMI 2025 · similarity 0.91)
- Global-Local Collaborative Inference with LLM for Lidar-Based Open-Vocabulary Detection (ECCV 2024 · similarity 0.91)
- Training an Open-Vocabulary Monocular 3D Detection Model without 3D Data (NeurIPS 2024 · similarity 0.91)
- 3D-AVS: LiDAR-based 3D Auto-Vocabulary Segmentation (CVPR 2025 · similarity 0.89)
- OLiDM: Object-aware LiDAR Diffusion Models for Autonomous Driving (AAAI 2025 · similarity 0.89)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
