Soiling detection for Advanced Driver Assistance Systems

TL;DR

This paper addresses soiling detection for automotive cameras by framing it as a semantic segmentation task, comparing methods, analyzing dataset issues, and providing a refined dataset for improved robustness in driver assistance systems.

Contribution

It introduces a new dataset subset for soiling detection, compares segmentation methods, and highlights dataset issues like data leakage and annotation inaccuracies.

Findings

Segmentation methods outperform tile-level classification.

The new dataset subset enables faster training with comparable results.

Original dataset contains data leakage and annotation errors.

Abstract

Soiling detection for automotive cameras is a crucial part of advanced driver assistance systems to make them more robust to external conditions like weather, dust, etc. In this paper, we regard the soiling detection as a semantic segmentation problem. We provide a comprehensive comparison of popular segmentation methods and show their superiority in performance while comparing them to tile-level classification approaches. Moreover, we present an extensive analysis of the Woodscape dataset showing that the original dataset contains a data-leakage and imprecise annotations. To address these problems, we create a new data subset, which, despite being much smaller, provides enough information for the segmentation method to reach comparable results in a much shorter time. All our codes and dataset splits are available at https://github.com/filipberanek/woodscape_revision.