Deep Semantic Segmentation for Automated Driving: Taxonomy, Roadmap and Challenges

TL;DR

This paper reviews semantic segmentation techniques for automated driving, highlighting challenges, proposing alternatives, and evaluating architectures on the CamVid dataset to improve accuracy and speed.

Contribution

It provides a taxonomy and roadmap for semantic segmentation in automated driving, addressing unique challenges and exploring alternative approaches.

Findings

Semantic segmentation algorithms need adaptation for automated driving.

Evaluation shows trade-offs between accuracy and speed of different architectures.

Challenges include ensuring safety, robustness, and real-time performance.

Abstract

Semantic segmentation was seen as a challenging computer vision problem few years ago. Due to recent advancements in deep learning, relatively accurate solutions are now possible for its use in automated driving. In this paper, the semantic segmentation problem is explored from the perspective of automated driving. Most of the current semantic segmentation algorithms are designed for generic images and do not incorporate prior structure and end goal for automated driving. First, the paper begins with a generic taxonomic survey of semantic segmentation algorithms and then discusses how it fits in the context of automated driving. Second, the particular challenges of deploying it into a safety system which needs high level of accuracy and robustness are listed. Third, different alternatives instead of using an independent semantic segmentation module are explored. Finally, an empirical evaluation of various semantic segmentation architectures was performed on CamVid dataset in terms of accuracy and speed. This paper is a preliminary shorter version of a more detailed survey which is work in progress.