Rethinking Geometric Depth in Monocular 3D Object Detection: A Projection-Consistent Reformulation
Organizations: Michigan State University · University of North Carolina at Chapel Hill
Abstract
Monocular 3D detection (Mono3D) hinges on recovering depth from a single image, for which the geometric formulation is widely adopted due to its simplicity and direct grounding in projective geometry. We show that this formulation harbors a systematic, projection-induced bias: substituting the 2D bounding-box height recovers the depth of an object's visible front surface rather than its center, causing size-dependent underestimation that persists even under ground-truth inputs. We address this front-surface bias with MonoPCA, a unified framework built on a projection-consistent reformulation. At its core, PCDepth retains the measurable and closes the gap to center depth through an orientation- and size-aware correction derived from perspective geometry, eliminating the bias without introducing ill-posed regression targets. Because PCDepth couples depth to predicted object size and orientation, attribute perception becomes essential. We therefore introduce GeoAlign, a teacher-student framework that transfers geometric representations from Vision Foundation Models into a lightweight student, strengthening attribute perception at no inference cost. Across four challenging benchmarks, MonoPCA achieves SoTA performance, including a +1.44% AP gain on the KITTI leaderboard and +2.80% on Omni3D. Code is available at https://github.com/alanzhangcs/MonoPCA.
Figures & tables
| Method | Extra Data | Test , AP | Test , AP | Val , AP | Val , AP | ||||||||
| Easy | Mod. | Hard | Easy | Mod. | Hard | Easy | Mod. | Hard | Easy | Mod. | Hard | ||
| OccupancyM3D Peng et al. (2024) | LiDAR | 25.55 | 17.02 | 14.79 | 35.38 | 24.18 | 21.37 | 26.87 | 19.96 | 17.15 | 35.72 | 26.60 | 23.68 |
| OPA-3D Su et al. (2023) | Depth | 24.68 | 17.17 | 14.14 | 32.50 | 23.14 | 20.30 | 24.97 | 19.40 | 16.59 | 33.80 | 25.51 | 22.13 |
| MonoTAKD Liu et al. (2025) | LiDAR | 27.91 | 19.43 | 16.51 | 38.75 | 27.76 | 24.14 | 34.36 | 22.61 | 19.88 | 42.86 | 29.41 | 26.47 |
| MonoCon Liu et al. (2022) | None | 22.50 | 16.46 | 13.95 | 31.12 | 22.10 | 19.00 | 26.33 | 19.01 | 15.98 | |||
| MonoDETR Zhang et al. (2023a) | 25.00 | 16.47 | 13.58 | 33.60 | 22.11 | 18.60 | 28.84 | 20.61 | 16.38 | 37.86 | 26.95 | 22.80 | |
| Method | IoU 0.7 | IoU 0.5 | ||||||
| AP | AP | AP | AP | |||||
| Easy | Mod. | Easy | Mod. | Easy | Mod. | Easy | Mod. | |
| GUP Net Lu et al. (2021) | 8.50 | 7.40 | 14.21 | 12.81 | 29.03 | 26.16 | 33.42 | 30.23 |
| DEVIANT Kumar et al. (2022) | 9.69 | 8.33 | 16.28 | 14.36 | 31.47 | 28.22 | 35.61 | 31.93 |
| MonoDETR Zhang et al. (2023a) | 9.53 | 8.19 | 16.39 | 14.41 | 31.81 | 28.35 | 35.70 | 31.96 |
| MonoDGP Pu et al. (2025) | 10.04 | 8.78 | 16.55 | 14.53 | 29.56 | 26.17 | 32.67 | 29.44 |
| Method | AP | Param(M) | GFLOPs(G) |
| MonoDGP Pu et al. (2025) | 18.72 | 38.90 | 69.0 |
| MonoCoP Zhang et al. (2026a) | 19.11 | 42.50 | 71.8 |
| MonoPCA | 20.55 | 42.00 | 62.5 |
| Method | Omni3D | KITTI Leaderboard | |||||||
| Easy | Mod. | Hard | |||||||
| MonoDGP Pu et al. (2025) | 26.35 | 18.72 | 15.97 | ||||||
| MonoCoP Zhang et al. (2026a) | 27.54 | 19.11 | 16.33 | ||||||
| ImVoxelNet Rukhovich et al. (2022) | 23.5 | 23.4 | 21.5 | 17.15 | 10.97 | 9.15 | |||
| SMOKE Liu et al. (2020) | 25.9 | 20.4 | 20.0 | 14.03 | 9.76 | 7.84 | |||
| Cube R-CNN Brazil et al. (2023) | 36.0 | 32.7 | 31.9 | 37.6 | 21.8 | 72.9 | 23.59 | 15.01 | 12.56 |
| (a) Components | (b) Depth Formulation | (c) MoB | (d) Alignment Loss | |||||
| PCDepth | GeoAlign | Mod. | Formulation | Mod. | Configuration | Mod. | Loss | Mod. |
| 23.98 | Direct Reg. | 23.82 | MoB (SA only) | 24.97 | None | 25.07 | ||
| 25.07 | GeoDepth ( ) | 24.73 | MoB (SM only) | 25.44 | MAE | 25.53 | ||
| 24.73 | GeoDepth ( ) | 24.18 | VFM only | 25.78 | MSE | 25.69 | ||
| 26.22 | PCDepth | 26.22 | MoB (full) | 26.22 | HFA | 26.22 | ||
| Method | Extra Data | Venue | Test , AP ( ) | Test , AP ( ) | ||||
| Easy | Mod. | Hard | Easy | Mod. | Hard | |||
| OccupancyM3D Peng et al. (2024) | LiDAR | CVPR 24 | 25.55 | 17.02 | 14.79 | 35.38 | 24.18 | 21.37 |
| OPA-3D Su et al. (2023) | Depth | ICRA 23 | 24.68 | 17.17 | 14.14 | 32.50 | 23.14 | 20.30 |
| MonoTAKD Liu et al. (2025) * | LiDAR | CVPR 25 | 27.91 | 19.43 | 16.51 | 38.75 | 27.76 | 24.14 |
| MonoFlex Zhang et al. (2021) | None | CVPR 21 | 19.94 | 13.89 | 12.07 | 28.23 | 19.75 | 16.89 |
| MonoRCNN Shi et al. (2021) | ICCV 21 | 18.36 | 12.65 | 10.03 | 25.48 | 18.11 | 14.10 | |
| Method | Extra Data | Venue | Val , AP ( ) | Val , AP ( ) | ||||
| Easy | Mod. | Hard | Easy | Mod. | Hard | |||
| OccupancyM3D Peng et al. (2024) | LiDAR | CVPR 24 | 26.87 | 19.96 | 17.15 | 35.72 | 26.60 | 23.68 |
| OPA-3D Su et al. (2023) | Depth | ICRA 23 | 24.97 | 19.40 | 16.59 | 33.80 | 25.51 | 22.13 |
| MonoTAKD Liu et al. (2025) * | LiDAR | CVPR 25 | 34.36 | 22.61 | 19.88 | 42.86 | 29.41 | 26.47 |
| MonoFlex Zhang et al. (2021) | None | CVPR 21 | 23.64 | 17.51 | 14.83 | |||
| MonoRCNN Shi et al. (2021) | ICCV 21 | |||||||
| Method | IoU 0.7 | IoU 0.5 | ||||||
| AP | AP | AP | AP | |||||
| Easy | Mod. | Easy | Mod. | Easy | Mod. | Easy | Mod. | |
| GUP Net Lu et al. (2021) | 8.50 | 7.40 | 14.21 | 12.81 | 29.03 | 26.16 | 33.42 | 30.23 |
| DEVIANT Kumar et al. (2022) | 9.69 | 8.33 | 16.28 | 14.36 | 31.47 | 28.22 | 35.61 | 31.93 |
| MonoDETR Zhang et al. (2023a) | 9.53 | 8.19 | 16.39 | 14.41 | 31.81 | 28.35 | 35.70 | 31.96 |
| MonoDGP Pu et al. (2025) | 10.04 | 8.78 | 16.55 | 14.53 | 29.56 | 26.17 | 32.67 | 29.44 |
| Method | Difficulty | APH (%) | AP (%) | ||||||
| All | 0-30 | 30-50 | 50- | All | 0-30 | 30-50 | 50- | ||
| GUP Net Lu et al. (2021) in Kumar et al. (2022) | Level 1 | 9.94 | 24.59 | 4.78 | 0.22 | 10.02 | 24.78 | 4.84 | 0.22 |
| DEVIANT Kumar et al. (2022) | 10.89 | 26.64 | 5.08 | 0.18 | 10.98 | 26.85 | 5.13 | 0.18 | |
| MonoDETR Zhang et al. (2023a) | 9.60 | 23.58 | 4.67 | 0.99 | 9.68 | 23.78 | 4.72 | 1.00 | |
| MonoUNI Jia et al. (2023) | 10.73 | 26.30 | 3.98 | 0.55 | 10.98 | 26.63 | 4.04 | 0.57 | |
| MonoDGP Pu et al. (2025) | 9.84 | 23.73 | 5.01 | 0.98 | 10.06 | 24.01 | 5.06 | 0.99 | |
| Method | Difficulty | APH (%) | AP (%) | ||||||
| All | 0-30 | 30-50 | 50- | All | 0-30 | 30-50 | 50- | ||
| GUP Net Lu et al. (2021) in Kumar et al. (2022) | Level 1 | 2.27 | 6.11 | 0.80 | 0.03 | 2.28 | 6.15 | 0.81 | 0.03 |
| DEVIANT Kumar et al. (2022) | 2.67 | 6.90 | 0.98 | 0.02 | 2.69 | 6.95 | 0.99 | 0.02 | |
| MonoDETR Zhang et al. (2023a) | 2.10 | 5.94 | 0.73 | 0.12 | 2.11 | 5.99 | 0.73 | 0.12 | |
| MonoUNI Jia et al. (2023) | 3.16 | 8.50 | 0.86 | 0.12 | 3.20 | 8.61 | 0.87 | 0.13 | |
| MonoDGP Pu et al. (2025) | 2.39 | 6.62 | 0.84 | 0.12 | 2.41 | 6.67 | 0.84 | 0.12 | |
| Method | AP on Category | |||||
| Cyclist | Truck | Pedestrian | Van | Car | Avg | |
| MonoPCA | 28.85 | 27.69 | 36.29 | 21.22 | 76.35 | 38.09 |
| MonoPCA- T | 27.68 | 27.25 | 38.23 | 28.62 | 76.18 | 39.60 |
| Method | AP on Category | |||||||||
| Pedestrian | Motorcycle | Truck | Trailer | Car | Traffic Cone | Bus | Barrier | Bicycle | Avg | |
| MonoPCA | 14.45 | 24.62 | 45.04 | 26.93 | 77.64 | 11.52 | 70.52 | 40.57 | 21.53 | 36.99 |
| MonoPCA- T | 15.15 | 27.62 | 44.41 | 27.39 | 78.93 | 17.18 | 67.98 | 51.04 | 23.01 | 39.20 |
| HFA Coefficient | Easy | Mod. | Hard |
| 1 | 33.46 | 25.06 | 21.90 |
| 2 | 34.73 | 25.94 | 22.12 |
| 5 | 35.86 | 26.22 | 23.58 |
| 10 | 33.89 | 25.21 | 21.91 |
Monocular 3D Object Detection
Monocular 3D object detection estimates objects’ positions, dimensions, and orientations in 3D from one camera, without depth sensors.
See all 24 papers