# Lagos Published articles for Lagos. This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles. ## Occlusion for Free: Detecting Occlusion from the Disagreement Between Two Monocular Distance Cues DevFeed: [Occlusion for Free: Detecting Occlusion from the Disagreement Between Two Monocular Distance Cues]() Original publisher: [Read original article]() Author: Chibueze Praise Onyenweaku, Toyibat Adele, Sikirulahi Opeyemi Abdulkareem, Obed Honour Eje Published: 2026-11-05T05:00:00Z Content type: article Language: en Sources: [arXiv: Artificial Intelligence]() Topics: [Computer vision](), [Machine learning]() Tags: [all](), [annotation](), [annotations](), [arxiv](), [attention](), [box](), [camera](), [computer-vision](), [cs-ai](), [cs-cv](), [detection](), [distribution](), [error](), [fusion](), [lagos](), [local](), [map](), [metric](), [model](), [ordering](), [pipeline](), [pixels]() ### AI overview The proposed training-free occlusion score compares apparent pedestrian height with monocular metric depth: occlusion largely preserves the box height but pulls the depth estimate toward the nearer occluder. It uses only the ordering of the two cues, requiring no focal length, assumed stature, depth scale, or occlusion labels, provided depth error is close to a pure gain--a condition the authors say can be checked without labels. It reports AUC 0.909 for heavy occlusion on KITTI (0.888 across all annotated pedestrians, including detector misses) and 0.819 on synthetic CARLA scenes, where distance alone reaches 0.644. Performance remains strong within fixed distance bands, suggesting the score is not merely a range proxy. The abstract also reports that, in tested configurations, depth fusion, occlusion-simulating augmentation, and attention did not improve detection outside their training distribution; a German-road-trained detector lost half its accuracy on self-annotated Lagos street scenes. ### Source excerpt arXiv:2611.06899v1 Announce Type: cross Abstract: 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.