Knowing that a pedestrian is occluded is useful to a robot even when the detector cannot recover the hidden pixels, but occlusion estimators are normally trained on occlusion annotations that deployment settings rarely have. We show that occlusion can be read off two distance estimates a camera-only pipeline already computes. Apparent height is largely unchanged by occlusion, because the detected box still spans the whole person; monocular metric depth is pulled toward the occluder, because the nearest surface inside the box is what a depth map reports. Only their ordering matters, so no focal length, stature, or depth scale is needed, provided the depth model's error is close to a pure gain, a condition that can be checked without labels. The score detects heavy occlusion on KITTI with AUC 0.909 (0.888 over all annotated pedestrians, including those the detector misses), requires no training and no occlusion labels, and is not a proxy for range: within fixed distance bands it holds AUC 0.894-0.925 while true distance alone sits near chance. On synthetic CARLA scenes with measured visibility it reaches 0.819, against 0.644 for distance alone. We arrived here after measuring three standard architectural responses to occlusion, depth fusion, occlusion-simulating augmentation, and attention, none of which improved detection outside its training distribution in the configurations we tested, and after finding that a detector trained on German roads loses half its accuracy on self-annotated street scenes in Lagos, Nigeria. Signals that need no local annotation are worth more in settings that benchmarks do not represent.
Figures & tables
Figure 1: Runtime pipeline. Offline inputs are fixed before deployment. At runtime, two monocular distance estimates per detection are compared; their disagreement flags occlusion without training or occlusion labels. LiDAR ground truth is used for evaluation only, and stature and focal length enter only metric read-outs, not the ranking score.
Figure 2: One KITTI frame showing the proposed cue. The visible pedestrian’s nearest surface is their own and the two distance estimates agree; for the occluded pedestrian, depth reads the nearer occluder and the estimates diverge by 34%.
Distance band
N vis/occ (peds)
AUC of score [95% CI]
AUC of distance
5–10 m
409/26 (137/10)
0.917 [0.843, 0.971]
0.594
10–20 m
466/140 (156/55)
0.894 [0.840, 0.939]
0.457
20 m +
318/100 (112/41)
0.925 [0.874, 0.964]
0.469
all
1237/269 (420/107)
0.909 [0.876, 0.938]
0.624
Table 1: Occlusion score on KITTI test, heavy occlusion vs. fully-visible. N counts matched detections, with distinct pedestrians in parentheses; intervals bootstrap pedestrians. The final column is the AUC of true distance alone on the same instances.
Appendix figures & tables8 assets
Supplementary material from the paper’s appendix.
Appendix
Target
Band
N vis/occ (peds)
AUC of score [95% CI]
AUC of distance
Any (1–3)
5–10 m
409/124 (137/44)
0.773 [0.676, 0.863]
0.549
Any (1–3)
10–20 m
466/297 (156/109)
0.807 [0.750, 0.860]
0.471
Any (1–3)
20 m +
318/267 (112/102)
0.874 [0.828, 0.919]
0.513
Any (1–3)
all
1237/704 (420/262)
0.826 [0.792, 0.857]
0.600
Partly (1)
5–10 m
409/98 (137/34)
0.735 [0.615, 0.840]
0.538
Partly (1)
10–20 m
466/157 (156/54)
0.728 [0.648, 0.809]
0.485
Appendix
Table 2: Occlusion score on KITTI test versus fully-visible pedestrians. N counts matched detections, with distinct pedestrians in parentheses; intervals bootstrap pedestrians. The final column is the AUC of true distance alone on the same instances.
Read-out
Calib.
AUC
Ranking, dminhpx
none
0.909
Signed, 1−zdep/zapp
fy,S,k
0.908
Absolute difference
fy,S,k
0.872
Absolute, gain 25% high
fy,S,k
0.470
Appendix
Table 3: Occlusion AUC for three formulations of the cue disagreement. Heavy occlusion.
Depth source
any
heavy
MiDaS, as-is
0.572
0.596
MiDaS, inverted
0.674
0.798
DAv2, metric
0.826
0.909
Appendix
Table 4: Occlusion AUC by depth source.
Variant
Heavy AUC [95% CI]
Paper score, dmin in predicted box
0.909 [0.876, 0.938]
h annotated, dmin predicted box
0.899 [0.866, 0.928]
h predicted, dmin annotated box
0.902 [0.868, 0.931]
Annotated boxes only, matched pedestrians
0.904 [0.870, 0.931]
Annotated boxes, all 709 pedestrians
0.888 [0.852, 0.921]
1st / 5th / 10th / 25th percentile instead of min
0.911 / 0.919 / 0.921 / 0.910
Appendix
Table 5: Mechanism checks, KITTI heavy occlusion vs. visible. Pedestrian-bootstrap 95% intervals.
Intervention
Setting
Δ mAP50
Δ Hard mAP50
Occlusion augmentation (GridMask)
KITTI, YOLOv8x
+0.009
+0.041
Occlusion augmentation (H&S)
CityPersons, YOLOv8x
+0.001
−0.004
Depth fusion (MiDaS)
KITTI, YOLOv8x
−0.023
−0.012
Depth fusion (MiDaS) †
KITTI, YOLOv8n
−0.075
−0.143
Depth fusion (DAv2) †
KITTI, YOLOv8n
−0.056
−0.093
Attention (POPAM) †
KITTI, YOLOv8n
−0.030
−0.055
Appendix
Table 6: Three model-side interventions against their RGB-only baselines. Pedestrian class; Δ is mAP50 relative to baseline. † denotes single-seed results, which record what that configuration did rather than a general conclusion.
Trained on
Configuration
In-dist. mAP50
Lagos mAP50
KITTI
Baseline
0.8809
0.4482 ± 0.0094
KITTI
GridMask
0.8903
0.4150 ± 0.0530
CityPersons
Baseline
0.7154
0.7356 ± 0.0351
CityPersons
Hide-and-Seek
0.7166
0.7075 ± 0.0198
Appendix
Table 7: Detector transfer to 350 self-annotated Lagos frames (1,290 pedestrian boxes), mean ± s.d. over three training seeds. No model has seen this data. The annotations carry no occlusion labels, so the occlusion score is not evaluated here.
Distance band
N
AUC of score [95% CI]
AUC of distance
0–5 m
154
0.969 [0.939, 0.989]
0.506
5–10 m
668
0.895 [0.868, 0.919]
0.585
10–15 m
687
0.839 [0.804, 0.871]
0.562
15–25 m
1277
0.797 [0.771, 0.821]
0.509
25–60 m
507
0.739 [0.693, 0.783]
0.481
all
3293
0.819 [0.801, 0.834]
0.644
Appendix
Table 8: Occlusion score on CARLA (visibility < 0.8 vs. ≥ 0.8), by distance band. N counts pedestrians; intervals bootstrap frames.
Depth input
Detection
Depth
Score
Total
FPS
518 × 1722 (default)
77
6233
0.3
6336
0.16
294 × 980
66
1229
0.3
1314
0.76
Appendix
Table 9: Median CPU latency per frame (ms), YOLOv8n + Depth Anything V2 Metric-Outdoor-Small.
Department of Computer and Information Science University of Michigan-Dearborn Dearborn, Michigan, USA · Department of Industrial and Manufacturing Systems Engineering University of Michigan-Dearborn Dearborn, Michigan, USA