Research map / Detection and segmentation / 3D detection and driving
MonoJSG: Joint Semantic and Geometric Cost Volume for Monocular 3D Object Detection
Qing Lian, Peiliang Li, Xiaozhi Chen
Check 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 pixel-level joint semantic and geometric cost-volume refinement | CenterNet-based monocular 3D detector · KITTI test Car, Moderate | 3D AP_R40: 14.17 -> 16.14 | Improved |
| Add pixel-level joint semantic and geometric cost-volume refinement | CenterNet-based monocular 3D detector · Waymo validation Car, Level 1, IoU 0.7, Overall | 3D mAP: 0.78 -> 0.97 | Improved |
| Replace sparse eight-corner geometric constraints with dense pixel-level geometric constraints (ablation) | MonoJSG depth recovery · KITTI validation Car, Easy | 3D AP: 18.9 -> 22.3 | Improved |
| Add learned semantic features to geometric constraints (ablation) | MonoJSG depth recovery · KITTI validation Car, Easy | 3D AP: 22.3 -> 24.3 | Improved |
| Add refinement-network processing of the joint cost volume (ablation) | MonoJSG Constraint · KITTI validation Car, Easy | 3D AP: 24.3 -> 26.4 | Improved |
| Add sparse geometric constraints from eight corners instead of the baseline (ablation) | MonoJSG depth recovery · KITTI validation Car, Easy | 3D AP: 20.5 -> 18.9 | Hurt |
| Add geometric coordinates to refinement features (ablation) | MonoJSG refinement · KITTI validation Car, Easy | 3D AP: 20.5 -> 22.4 | Improved |
| Add raw photo features to geometric refinement features (ablation) | MonoJSG refinement · KITTI validation Car, Easy | 3D AP: 22.4 -> 24.3 | Improved |
| Add learned semantic features to geometric refinement features (ablation) | MonoJSG refinement · KITTI validation Car, Easy | 3D AP: 22.4 -> 26.3 | Improved |
| Add raw photo features to geometric and semantic refinement features (ablation) | MonoJSG refinement · KITTI validation Car, Easy | 3D AP: 26.3 -> 26.4 | No clear effect |
| Use adaptive rather than uniform depth sampling with sample size 8 (ablation) | MonoJSG cost volume · KITTI validation Car, Easy | 3D AP: 23.6 -> 24.7 | Improved |
| Use adaptive rather than uniform depth sampling with sample size 32 (ablation) | MonoJSG cost volume · KITTI validation Car, Easy | 3D AP: 25.2 -> 26.4 | Improved |
Later papers that compared with this method
2 later accepted papers report MonoJSG as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.
- FD3D: Exploiting Foreground Depth Map for Feature-Supervised Monocular 3D Object Detection (AAAI 2024 · better on 1 of 1 shared result)
- DID-M3D: Decoupling Instance Depth for Monocular 3D Object Detection (ECCV 2022 · better on 1 of 1 shared result)
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
- MonoDGP: Monocular 3D Object Detection with Decoupled-Query and Geometry-Error Priors (CVPR 2025 · similarity 0.90)
- RefineAny3D: Depth Refinement as Semantic Alignment for Monocular 3D Detection (NeurIPS 2026 · similarity 0.90)
- Rethinking Geometric Depth in Monocular 3D Object Detection: A Projection-Consistent Reformulation (NeurIPS 2026 · similarity 0.90)
- MonoCD: Monocular 3D Object Detection with Complementary Depths (CVPR 2024 · similarity 0.89)
- MonoGRNet: A General Framework for Monocular 3D Object Detection (TPAMI 2021 · similarity 0.89)
- MonoMAE: Enhancing Monocular 3D Detection through Depth-Aware Masked Autoencoders (NeurIPS 2024 · similarity 0.88)
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.
