Research map / Point clouds / Registration and matching
TrafficLoc: Localizing Traffic Surveillance Cameras in 3D Scenes
Yan Xia, Yunxiang Lu, Rui Song, Oussema Dhaouadi, João F. Henriques, Daniel Cremers
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 Geometry-guided Attention Loss (ablation) | TrafficLoc · Carla Intersection Test T1-T7 | RTE (m): 0.62 -> 0.51 | Improved |
| add GAL (ablation) | TrafficLoc · Carla Intersection Test T1-T7 hard | RTE (m): 1.42 -> 1.13 | Improved |
| use Inter-intra Contrastive Learning instead of normal contrastive learning (ablation) | TrafficLoc · Carla Intersection Test T1-T7 | RTE (m): 0.82 -> 0.74 | Improved |
| use ICL instead of NCL (ablation) | TrafficLoc · Carla Intersection Test T10 | RRE (deg): 7.35 -> 4.09 | Improved |
| add Dense Training Alignment (ablation) | TrafficLoc · Carla Intersection Test T1-T7 | RRE (deg): 1.27 -> 1.01 | Improved |
| add fine matching (ablation) | TrafficLoc · Carla Intersection Test T1-T7 | RRE (deg): 1.01 -> 0.84 | Improved |
| set equal lower and upper GAL thresholds (10 deg, 10 deg) (ablation) | TrafficLoc · Carla Intersection Test T1-T7 | RRE (deg): 0.84 -> 1.24 | Hurt |
| apply GAL to the first fusion layer instead of the last (ablation) | TrafficLoc · Carla Intersection Test T1-T7 | RRE (deg): 0.66 -> 1.00 | Hurt |
| apply GAL to all fusion layers instead of the last (ablation) | TrafficLoc · Carla Intersection Test T1-T7 | RRE (deg): 0.66 -> 1.02 | Hurt |
| add DTA and GAL to CoFiI2P (ablation) | CoFiI2P · KITTI Odometry | RRE (deg): 1.14 -> 0.85 | Improved |
| use DUSt3R instead of ResNet image encoder (ablation) | TrafficLoc · Carla Intersection Test T1-T7 | RRE (deg): 0.85 -> 0.66 | Improved |
| use four fusion layers instead of two (ablation) | TrafficLoc · Carla Intersection Test T1-T7 | RRE (deg): 0.96 -> 0.66 | Improved |
Later papers that compared with this method
1 later accepted paper reports TrafficLoc as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- DRS-VPT: Directly Re-localizing in Scenes using a Vision and Point Transformer (ECCV 2026 · better on 4 of 4 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
- Efficient Global 2D-3D Matching for Camera Localization in a Large-Scale 3D Map (ICCV 2017 · similarity 0.87)
- Sparse Point Guided 3D Lane Detection (ICCV 2023 · similarity 0.84)
- DV-3DLane: End-to-end Multi-modal 3D Lane Detection with Dual-view Representation (ICLR 2024 · similarity 0.84)
- 3D StreetUnveiler with Semantic-aware 2DGS - a simple baseline (ICLR 2025 · similarity 0.84)
- Sketch and Refine: Towards Fast and Accurate Lane Detection (AAAI 2024 · similarity 0.84)
- Traffic Scene Parsing through the TSP6K Dataset (CVPR 2024 · similarity 0.83)
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.
