Research map / Detection and segmentation / 3D detection and driving
PerspectiveNet: 3D Object Detection from a Single RGB Image via Perspective Points
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 PerspectiveNet with ground-truth camera extrinsic parameters instead of CooP | PerspectiveNet · SUN RGB-D | mAP: 23.65 -> 39.09 | Helped |
| Use ground-truth camera extrinsic parameters instead of learning them jointly | PerspectiveNet · SUN RGB-D | mAP: 34.96 -> 39.09 | Helped |
| Add the perspective point branch instead of predicting the 3D center from the 2D offset (ablation) | PerspectiveNet · SUN RGB-D | mAP: 35.23 -> 39.09 | Helped |
| Use gated feature aggregation instead of direct aggregation (ablation) | PerspectiveNet · SUN RGB-D | mAP: 38.63 -> 39.09 | Helped |
| Use a soft gate instead of a hard 0-or-1 gate (ablation) | PerspectiveNet · SUN RGB-D | mAP: 38.87 -> 39.09 | Helped |
| Add the perspective loss (ablation) | PerspectiveNet · SUN RGB-D | mAP: 39.01 -> 39.09 | Helped |
| Add the 2D-3D consistency loss (ablation) | PerspectiveNet · SUN RGB-D | mAP: 37.43 -> 39.09 | Helped |
| Add the perspective branch, perspective loss, and consistency loss (ablation) | PerspectiveNet · SUN RGB-D | mAP: 32.97 -> 39.09 | Helped |
| Use template-based rather than heatmap-based perspective-point estimation (ablation) | PerspectiveNet perspective-point estimator · SUN RGB-D | average perspective-point absolute distance (pix): 10.25 pix -> 6.37 pix | Helped |
Closest papers by meaning
- 3D Bounding Box Estimation Using Deep Learning and Geometry (CVPR 2017 · similarity 0.90)
- Towards Generalizable Multi-Camera 3D Object Detection via Perspective Rendering (AAAI 2025 · similarity 0.89)
- To the Point: Efficient 3D Object Detection in the Range Image With Graph Convolution Kernels (CVPR 2021 · similarity 0.89)
- ImageNet3D: Towards General-Purpose Object-Level 3D Understanding (NeurIPS 2024 · similarity 0.88)
- MonoDGP: Monocular 3D Object Detection with Decoupled-Query and Geometry-Error Priors (CVPR 2025 · similarity 0.87)
- Alignment-Free RGB-T Salient Object Detection: A Large-Scale Dataset and Progressive Correlation Network (AAAI 2025 · similarity 0.86)
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.