# Unaligned but Safe -- Formally Compensating Performance Limitations for Imprecise 2D Object Detection

arXiv: https://arxiv.org/abs/2202.05123 · HTML: https://mlsanity.com/paper/2202.05123/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Builds on 8 of its 19 references that are in the map.
- First posted on arXiv 10 Feb 2022.

## Builds on (most cited first, 8 of 8)

- [Microsoft COCO: Common Objects in Context](https://mlsanity.com/paper/1405.0312/) (2014)
- [DeepXplore: Automated Whitebox Testing of Deep Learning Systems](https://mlsanity.com/paper/1705.06640/) (2017)
- [Inspect, Understand, Overcome: A Survey of Practical Methods for AI Safety](https://mlsanity.com/paper/2104.14235/) (2021)
- [A Safety Framework for Critical Systems Utilising Deep Neural Networks](https://mlsanity.com/paper/2003.05311/) (2020)
- [Quantitative Projection Coverage for Testing ML-enabled Autonomous Systems](https://mlsanity.com/paper/1805.04333/) (2018)
- [The missing link: Developing a safety case for perception components in automated driving](https://mlsanity.com/paper/2108.13294/) (2021)
- [A Framework for Assurance of Medication Safety using Machine Learning](https://mlsanity.com/paper/2101.05620/) (2021)
- [Logically Sound Arguments for the Effectiveness of ML Safety Measures](https://mlsanity.com/paper/2111.02649/) (2021)

## Built on by (most cited first, 3 of 3)

- [USC: Uncompromising Spatial Constraints for Safety-Oriented 3D Object Detectors in Autonomous Driving](https://mlsanity.com/paper/2209.10368/) (2022)
- [EC-IoU: Orienting Safety for Object Detectors via Ego-Centric Intersection-over-Union](https://mlsanity.com/paper/2403.15474/) (2024)
- [Architecting Safer Autonomous Aviation Systems](https://mlsanity.com/paper/2301.08138/) (2023)

## Abstract (arXiv)

In this paper, we consider the imperfection within machine learning-based 2D object detection and its impact on safety. We address a special sub-type of performance limitations: the prediction bounding box cannot be perfectly aligned with the ground truth, but the computed Intersection-over-Union metric is always larger than a given threshold. Under such type of performance limitation, we formally prove the minimum required bounding box enlargement factor to cover the ground truth. We then demonstrate that the factor can be mathematically adjusted to a smaller value, provided that the motion planner takes a fixed-length buffer in making its decisions. Finally, observing the difference between an empirically measured enlargement factor and our formally derived worst-case enlargement factor offers an interesting connection between the quantitative evidence (demonstrated by statistics) and the qualitative evidence (demonstrated by worst-case analysis).