Research map / Detection and segmentation / 3D detection and driving
DDStereo: Efficient Dual Decoder Transformers for Stereo 3D Road Anomaly Detection
Shiyi Mu, Zichong Gu, Zhiqi Ai, Yilin Gao, Shugong Xu
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 |
|---|---|---|---|
| use shared queries instead of independent queries (ablation) | DDStereo · AR-OoD | OoD AP_3D: 51.19 -> 78.09 | Improved |
| use shared queries instead of independent queries (ablation) | DDStereo · KITTI val Car | AP_3D moderate: 44.3 -> 48.2 | Improved |
| use dual decoder instead of single decoder (ablation) | DDStereo · AR-OoD | OoD AP_3D: 66.37 -> 78.09 | Improved |
| use grid sample depth instead of average strategy (ablation) | DDStereo · AR-OoD | OoD AP_3D: 69.39 -> 78.09 | Improved |
| use MNPF (sigmoid) score instead of MaxLogit (ablation) | DDStereo · AR-OoD | OoD AP_3D: 67.71 -> 78.09 | Improved |
| use MNPF score instead of MSP (ablation) | DDStereo · AR-OoD | OoD AP_3D: 75.16 -> 78.09 | Improved |
| DDStereo vs S3AD with 10% extra synthetic data | DDStereo · AR-OoD | OoD AP_3D: 24.45 -> 54.65 | Improved |
| DDStereo vs StereoDETR (adds open-set capability) | DDStereo · KITTI test Car | AP_3D moderate / latency: 41.17 -> 43.97; 18 ms -> 24 ms | Improved |
| DDStereo vs S3AD on OoD R11 AP_3D | DDStereo · AR-OoD | OoD AP_3D (R11): 74.35 vs 74.23 | No clear effect |
| DDStereo on KITTI Easy / BEV | DDStereo · KITTI test Car | AP_BEV moderate: 54.53 (StereoDETR) -> 53.60 | Hurt |
Closest papers by meaning
- OpenAD: Open-World Autonomous Driving Benchmark for 3D Object Detection (NeurIPS 2025 · similarity 0.86)
- MonoDETR: Depth-guided Transformer for Monocular 3D Object Detection (ICCV 2023 · similarity 0.86)
- MonoDGP: Monocular 3D Object Detection with Decoupled-Query and Geometry-Error Priors (CVPR 2025 · similarity 0.85)
- Spotting the Unexpected (STU): A 3D LiDAR Dataset for Anomaly Segmentation in Autonomous Driving (CVPR 2025 · similarity 0.85)
- DepthVanish: Optimizing Adversarial Interval Structures for Stereo-Depth-Invisible Patches (NeurIPS 2025 · similarity 0.84)
- Any-Stereo: Arbitrary Scale Disparity Estimation for Iterative Stereo Matching (AAAI 2024 · similarity 0.84)
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.
