Research map / Detection and segmentation / 3D detection and driving
Learned Two-Plane Perspective Prior Based Image Resampling for Efficient Object Detection
Anurag Ghosh, N. Dinesh Reddy, Christoph Mertz, Srinivasa G. Narasimhan
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 |
|---|---|---|---|
| Replace uniform downsampling with learned two-plane perspective resampling | Faster R-CNN · Argoverse-HD at 0.5x | AP_S: +9.6 AP_S | Improved |
| Replace Fovea temporal sampling with learned two-plane perspective resampling | Faster R-CNN · Argoverse-HD at 0.5x | AP_S: +4.2 AP_S | Improved |
| Use Two-Plane Prior instead of Fovea temporal sampling | Faster R-CNN streaming · Argoverse-HD | sAP_S: +5.3 sAP_S | Improved |
| Use learned two-plane resampling instead of uniform downsampling | Faster R-CNN · WALT at 0.375x | AP_S: +10.7 AP_S | Improved |
| Use learned two-plane resampling instead of Fovea | Faster R-CNN · WALT at 0.375x | AP_S: +3.0 AP_S | Improved |
| Use learned two-plane resampling instead of uniform downsampling | Faster R-CNN · Commuter Bus at 0.5x | AR: 42% | Improved |
| Add an upper plane to the ground-plane-only prior (ablation) | Two-Plane Prior · Argoverse-HD | AP: +1.8 AP | Improved |
| Add an upper plane to the ground-plane-only prior (ablation) | Two-Plane Prior · Argoverse-HD | AP_S: Ground Plane Prior outperforms Two-Plane Prior on small objects by +1 AP_S | Hurt |
| Add an upper plane to the ground-plane-only prior (ablation) | Two-Plane Prior · Argoverse-HD | AP_M: +4.1 AP_M | Improved |
| Add an upper plane to the ground-plane-only prior (ablation) | Two-Plane Prior · Argoverse-HD | AP_L: +4.6 AP_L | Improved |
| Estimate the vanishing point rather than use an average point (ablation) | Two-Plane Prior · Argoverse-HD | AP: +1.2 AP | Improved |
| Train the Two-Plane Prior with pseudo-labels instead of ground-truth labels (ablation) | Two-Plane Prior · Argoverse-HD | AP: -3.7 AP | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- Faster R-CNN — Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks (TPAMI 2016) · better on 6 of 6 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
- Real-time 3D Object Detection with Inference-Aligned Learning (AAAI 2026 · similarity 0.86)
- ZoomNet: Part-Aware Adaptive Zooming Neural Network for 3D Object Detection (AAAI 2020 · similarity 0.86)
- Rethinking Geometric Depth in Monocular 3D Object Detection: A Projection-Consistent Reformulation (NeurIPS 2026 · similarity 0.85)
- DeNet: Scalable Real-Time Object Detection With Directed Sparse Sampling (ICCV 2017 · similarity 0.85)
- Towards Generalizable Multi-Camera 3D Object Detection via Perspective Rendering (AAAI 2025 · similarity 0.85)
- BEVSpread: Spread Voxel Pooling for Bird's-Eye-View Representation in Vision-based Roadside 3D Object Detection (CVPR 2024 · similarity 0.85)
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.
