Research map / Detection and segmentation / 3D detection and driving
The ApolloScape Open Dataset for Autonomous Driving and Its Application
Xinyu Huang, Peng Wang, Xinjing Cheng, Dingfu Zhou, Qichuan Geng, Ruigang Yang
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 semantic cues (road prior and semantic weights) to pose CNN (ablation) | Pose CNN · Zpark pose | rotation error (deg): from 0.982 to 0.727 | Improved |
| add semantic cues to pose CNN (ablation) | Pose CNN · Zpark pose | translation error (m): 1.355 -> 1.331 | Improved |
| add rendered label map with projected labels as input to pose CNN | noisy GPS/IMU pose · Zpark pose | translation / rotation error: 3.45 -> 1.355 m; 7.87 -> 0.982 deg | Improved |
| add RNN over pose CNN output (CNN-RNN) instead of pose CNN alone (ablation) | Pose CNN · Zpark pose | translation error (m): 1.331 -> 1.005 | Improved |
| add constant-speed Kalman filter on pose CNN output (ablation) | Pose CNN · Zpark pose | rotation error (deg): rotation 0.727 -> 0.833; translation 1.331 -> 1.281 | Mixed |
| add pose CNN in front of RNN instead of RNN on GPS/IMU alone (ablation) | Pose RNN · Zpark pose | rotation error (deg): 1.731 -> 0.719 | Improved |
| add pose CNN in front of RNN on Dlake (ablation) | Pose RNN · Dlake pose | translation error (m): 1.385 -> 0.890 | Improved |
| add rendered label map with pose RNN as guidance to segment CNN (ablation) | Segment CNN · Zpark segmentation | mIoU: 68.35 -> 69.93 | Improved |
| add rendered label map with pose CNN output as guidance to segment CNN (ablation) | Segment CNN · Zpark segmentation | mIoU: 68.35 -> 68.6 | No clear effect |
| add rendered label map with pose RNN guidance on larger Dlake (ablation) | Segment CNN · Dlake segmentation | mIoU: from 62.36 to 67.00 | Improved |
| use ground-truth pose for rendered label guidance (upper bound) (ablation) | Segment CNN · Zpark segmentation | mIoU: 68.35 -> 79.37 | Improved |
| use rendered label map alone as segmentation (no segment CNN) (ablation) | Render PoseRNN · Zpark segmentation | mIoU: 68.35 -> 32.61 | Hurt |
Closest papers by meaning
- ApolloCar3D: A Large 3D Car Instance Understanding Benchmark for Autonomous Driving (CVPR 2019 · similarity 0.90)
- One Million Scenes for Autonomous Driving: ONCE Dataset (NeurIPS 2021 · similarity 0.89)
- Scalability in Perception for Autonomous Driving: Waymo Open Dataset (CVPR 2020 · similarity 0.88)
- SelfD: Self-Learning Large-Scale Driving Policies From the Web (CVPR 2022 · similarity 0.88)
- The Road Ahead in Autonomous Driving: The KITScenes Multimodal Dataset (NeurIPS 2026 · similarity 0.87)
- OmniHD-Scenes: A Next-Generation Multimodal Dataset for Autonomous Driving (TPAMI 2026 · similarity 0.87)
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.
