Research map / Detection and segmentation / 3D detection and driving
3D Box Proposals From a Single Monocular Image of an Indoor Scene
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 |
|---|---|---|---|
| replace two-stage pipeline with end-to-end residual volumetric model | Est-DSS · NYUv2 | overall recall (%): 66.1 -> 69.3 | Helped |
| replace two-stage pipeline with end-to-end residual volumetric model | Est-DSS · NYUv2 | ABO: 0.348 -> 0.364 | Helped |
| use projective approximate TSDF instead of accurate TSDF (ablation) | Est-DSS · NYUv2 | overall recall (%): 66.1 -> 63.6 | Hurt |
| volumetric representation instead of direct 3D regression (Faster R-CNN baseline) | Est-DSS · NYUv2 | overall recall (%): 62.3 -> 66.1 | Helped |
| add residual path to compensate approximate TSDF (ablation) | Est-DSS-Approx · NYUv2 | overall recall (%): 63.6 -> 69.3 | Helped |
| increase depth loss weight from 1 to 10 (ablation) | Ours · NYUv2 | overall recall (%): 69.3 -> 67.4 | Hurt |
| increase depth loss weight from 1 to 10 (ablation) | Ours · NYUv2 | ABO: 0.364 -> 0.359 | Hurt |
| use only monocular RGB instead of ground-truth depth (vs 3D Selective Search) | Ours · NYUv2 | recall (%): 74.2 -> 69.3 | Hurt |
| use only monocular RGB instead of ground-truth depth (vs DSS) | Ours · NYUv2 | recall (%): 84.9 -> 69.3 | Hurt |
| evaluate on SUN-RGBD without fine-tuning, ours vs Est-DSS | Ours · SUN-RGBD excluding NYUv2 | recall (%): 64.1 -> 67.9 | Helped |
| evaluate on SUN-RGBD without fine-tuning, ours vs Est-DSS | Ours · SUN-RGBD excluding NYUv2 | ABO: 0.328 -> 0.353 | Helped |
Closest papers by meaning
- MonoCD: Monocular 3D Object Detection with Complementary Depths (CVPR 2024 · similarity 0.86)
- VSRD: Instance-Aware Volumetric Silhouette Rendering for Weakly Supervised 3D Object Detection (CVPR 2024 · similarity 0.86)
- MonoDGP: Monocular 3D Object Detection with Decoupled-Query and Geometry-Error Priors (CVPR 2025 · similarity 0.86)
- MonoPlace3D: Learning 3D-Aware Object Placement for 3D Monocular Detection (CVPR 2025 · similarity 0.86)
- Cubify Anything: Scaling Indoor 3D Object Detection (CVPR 2025 · similarity 0.86)
- FD3D: Exploiting Foreground Depth Map for Feature-Supervised Monocular 3D Object Detection (AAAI 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.