False Negative Reduction in Semantic Segmentation under Domain Shift using Depth Estimation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Maag, Kira, Rottmann, Matthias
Format: Preprint
Published: 2022
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918281655353344
author Maag, Kira
Rottmann, Matthias
author_facet Maag, Kira
Rottmann, Matthias
contents State-of-the-art deep neural networks demonstrate outstanding performance in semantic segmentation. However, their performance is tied to the domain represented by the training data. Open world scenarios cause inaccurate predictions which is hazardous in safety relevant applications like automated driving. In this work, we enhance semantic segmentation predictions using monocular depth estimation to improve segmentation by reducing the occurrence of non-detected objects in presence of domain shift. To this end, we infer a depth heatmap via a modified segmentation network which generates foreground-background masks, operating in parallel to a given semantic segmentation network. Both segmentation masks are aggregated with a focus on foreground classes (here road users) to reduce false negatives. To also reduce the occurrence of false positives, we apply a pruning based on uncertainty estimates. Our approach is modular in a sense that it post-processes the output of any semantic segmentation network. In our experiments, we observe less non-detected objects of most important classes and an enhanced generalization to other domains compared to the basic semantic segmentation prediction.
format Preprint
id arxiv_https___arxiv_org_abs_2207_03513
institution arXiv
publishDate 2022
record_format arxiv
spellingShingle False Negative Reduction in Semantic Segmentation under Domain Shift using Depth Estimation
Maag, Kira
Rottmann, Matthias
Computer Vision and Pattern Recognition
State-of-the-art deep neural networks demonstrate outstanding performance in semantic segmentation. However, their performance is tied to the domain represented by the training data. Open world scenarios cause inaccurate predictions which is hazardous in safety relevant applications like automated driving. In this work, we enhance semantic segmentation predictions using monocular depth estimation to improve segmentation by reducing the occurrence of non-detected objects in presence of domain shift. To this end, we infer a depth heatmap via a modified segmentation network which generates foreground-background masks, operating in parallel to a given semantic segmentation network. Both segmentation masks are aggregated with a focus on foreground classes (here road users) to reduce false negatives. To also reduce the occurrence of false positives, we apply a pruning based on uncertainty estimates. Our approach is modular in a sense that it post-processes the output of any semantic segmentation network. In our experiments, we observe less non-detected objects of most important classes and an enhanced generalization to other domains compared to the basic semantic segmentation prediction.
title False Negative Reduction in Semantic Segmentation under Domain Shift using Depth Estimation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2207.03513