# Efficacy of Pixel-Level OOD Detection for Semantic Segmentation

arXiv: https://arxiv.org/abs/1911.02897 · HTML: https://mlsanity.com/paper/1911.02897/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 7 Nov 2019.

## Built on by (most cited first, 9 of 9)

- [SegmentMeIfYouCan: A Benchmark for Anomaly Segmentation](https://mlsanity.com/paper/2104.14812/) (2021)
- [Unmasking Anomalies in Road-Scene Segmentation](https://mlsanity.com/paper/2307.13316/) (2023)
- [Inspect, Understand, Overcome: A Survey of Practical Methods for AI Safety](https://mlsanity.com/paper/2104.14235/) (2021)
- [Two Video Data Sets for Tracking and Retrieval of Out of Distribution Objects](https://mlsanity.com/paper/2210.02074/) (2022)
- [Dense outlier detection and open-set recognition based on training with noisy negative images](https://mlsanity.com/paper/2101.09193/) (2021)
- [Out-of-Distribution Detection for LiDAR-based 3D Object Detection](https://mlsanity.com/paper/2209.14435/) (2022)
- [Towards Training-free Open-world Segmentation via Image Prompt Foundation Models](https://mlsanity.com/paper/2310.10912/) (2023)
- [Detecting and Learning the Unknown in Semantic Segmentation](https://mlsanity.com/paper/2202.08700/) (2022)
- [Understanding the Impact of Edge Cases from Occluded Pedestrians for ML Systems](https://mlsanity.com/paper/2204.12402/) (2022)

## Abstract (arXiv)

The detection of out of distribution samples for image classification has been widely researched. Safety critical applications, such as autonomous driving, would benefit from the ability to localise the unusual objects causing the image to be out of distribution. This paper adapts state-of-the-art methods for detecting out of distribution images for image classification to the new task of detecting out of distribution pixels, which can localise the unusual objects. It further experimentally compares the adapted methods on two new datasets derived from existing semantic segmentation datasets using PSPNet and DeeplabV3+ architectures, as well as proposing a new metric for the task. The evaluation shows that the performance ranking of the compared methods does not transfer to the new task and every method performs significantly worse than their image-level counterparts.