mlsanityData to 7 Oct 2026

Depth as Prior Knowledge for Object Detection

arXiv 2602.05730 ↗First posted 5 Feb 2026cs.CVcited by 0

Not on the map. Why?

Abstract (arXiv)

Detecting small and distant objects remains challenging for object detectors due to scale variation, low resolution, and background clutter. Safety-critical applications require reliable detection of these objects for safe planning. Depth information can improve detection, but existing approaches require complex, model-specific architectural modifications. We provide a theoretical analysis followed by an empirical investigation of the depth-detection relationship. Together, they explain how depth causes systematic performance degradation and why depth-informed supervision mitigates it. We introduce DepthPrior, a framework that uses depth as prior knowledge rather than as a fused feature, providing comparable benefits without modifying detector architectures. DepthPrior consists of Depth-Based Loss Weighting (DLW) and Depth-Based Loss Stratification (DLS) during training, and Depth-Aware Confidence Thresholding (DCT) during inference. The only overhead is the initial cost of depth estimation. Experiments across four benchmarks (KITTI, MS COCO, VisDrone, SUN RGB-D) and two detectors (YOLOv11, EfficientDet) demonstrate the effectiveness of DepthPrior, achieving up to +9% mAPS and +7% mARS for small objects, with inference recovery rates as high as 95:1 (true vs. false detections). DepthPrior offers these benefits without additional sensors, architectural changes, or performance costs. Code is available at https://github.com/mos-ks/DepthPrior.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 29 of its 46 references that are in the map. First posted on arXiv 5 Feb 2026.

Builds on

29 of 46 references.

References · 29

  1. 2014Microsoft COCO: Common Objects in Contextcited by 17,498
  2. 2017On Calibration of Modern Neural Networkscited by 4,860
  3. 2017What Uncertainties Do We Need in Bayesian Deep Learning for Computer Vision?cited by 2,335
  4. 2021Vision Transformers for Dense Predictioncited by 1,789
  5. 2024Depth Anything: Unleashing the Power of Large-Scale Unlabeled Datacited by 1,346
  6. 2017Multi-Task Learning Using Uncertainty to Weigh Losses for Scene Geometry and Semanticscited by 1,345
  7. 2014Depth Map Prediction from a Single Image using a Multi-Scale Deep Networkcited by 1,264
  8. 2019EfficientDet: Scalable and Efficient Object Detectioncited by 1,018
  9. 2017A Closer Look at Memorization in Deep Networkscited by 936
  10. 2018An Empirical Study of Example Forgetting during Deep Neural Network Learningcited by 669
  11. 2016Training Region-based Object Detectors with Online Hard Example Miningcited by 577
  12. 2018Pseudo-LiDAR from Visual Depth Estimation: Bridging the Gap in 3D Object Detection for Autonomous Drivingcited by 455
17 more
  1. 2019Which Tasks Should Be Learned Together in Multi-task Learning?cited by 367
  2. 2018Not All Samples Are Created Equal: Deep Learning with Importance Samplingcited by 323
  3. 2021Objects are Different: Flexible Monocular 3D Object Detectioncited by 184
  4. 2019Learning Depth-Guided Convolutions for Monocular 3D Object Detectioncited by 169
  5. 2020MonoPair: Monocular 3D Object Detection Using Pairwise Spatial Relationshipscited by 153
  6. 2014Learning Rich Features from RGB-D Images for Object Detection and Segmentationcited by 124
  7. 2017Learning non-maximum suppressioncited by 112
  8. 2018Auxiliary Tasks in Multi-task Learningcited by 106
  9. 2022MonoDETR: Depth-guided Transformer for Monocular 3D Object Detectioncited by 85
  10. 2016Marr Revisited: 2D-3D Alignment via Surface Normal Predictioncited by 34
  11. 2018Depth CNNs for RGB-D scene recognition: learning from scratch better than transferring from RGB-CNNscited by 7
  12. 2022Prior Knowledge-Guided Attention in Self-Supervised Vision Transformerscited by 7
  13. 2022Understanding the Impact of Image Quality and Distance of Objects to Object Detection Performancecited by 4
  14. 2024Cost-Sensitive Uncertainty-Based Failure Recognition for Object Detectioncited by 3
  15. 2024Self-Supervised Learning for Real-World Object Detection: a Surveycited by 2
  16. 2016Deeply Exploit Depth Information for Object Detectioncited by 1
  17. 2022Does depth estimation help object detection?cited by 1

Built on by

Not cited yet.