Research map / Point clouds / Point cloud classification
NeurVPS: Neural Vanishing Point Scanning via Conic Convolution
Yichao Zhou, Haozhi Qi, Jingwei Huang, Yi Ma
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 conic convolutions with regular convs and regress line direction (REG) (ablation) | NeurVPS · SU3 wireframe | AA1.0: 15.0 -> 86.3 | Improved |
| add conic convolution instead of regular convs with coordinate input (CLS) (ablation) | CLS classifier · SU3 wireframe | AA1.0: 23.7 -> 86.3 | Improved |
| increase conic conv layers from 2 to 4 (ablation) | NeurVPS · SU3 wireframe | AA0.2: 10.5 -> 47.5 | Improved |
| increase conic conv layers from 4 to 6 (ablation) | NeurVPS · SU3 wireframe | AA0.2: 47.5 -> 49.1 | No clear effect |
| use NeurVPS (conic conv) instead of vpdet on Natural Scene | vpdet · Natural Scene | AA1: 18.5 -> 29.1 | Improved |
| use NeurVPS instead of regular-conv CLS on ScanNet (ablation) | CLS · ScanNet | AA1: 2.0 -> 3.4 | Improved |
| use neural network instead of LSD + J-linkage | LSD + J-linkage · SU3 wireframe | AA0.2: 27.9 -> 2.2 (REG) | Mixed |
| NeurVPS vs LSD + J-linkage on ScanNet | LSD + J-linkage · ScanNet | mean angle error: 12.6◦ -> 4.5◦ | Improved |
Closest papers by meaning
- Deep Vanishing Point Detection: Geometric Priors Make Dataset Variations Vanish (CVPR 2022 · similarity 0.91)
- VaPiD: A Rapid Vanishing Point Detector via Learned Optimizers (ICCV 2021 · similarity 0.88)
- Detecting Vanishing Points Using Global Image Context in a Non-Manhattan World (CVPR 2016 · similarity 0.88)
- End-to-End Real-Time Vanishing Point Detection with Transformer (AAAI 2024 · similarity 0.87)
- VPDETR: End-to-End Vanishing Point DEtection TRansformers (AAAI 2024 · similarity 0.86)
- DeCoTR: Enhancing Depth Completion with 2D and 3D Attentions (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.
