Domain-Incremental Semantic Segmentation for Autonomous Driving under Adverse Driving Conditions

TL;DR

This paper introduces Progressive Semantic Segmentation (PSS), a domain-incremental learning architecture for autonomous driving that maintains performance across diverse adverse conditions without catastrophic forgetting.

Contribution

The paper presents a novel task-agnostic, dynamically growing collection of domain-specific models for semantic segmentation, addressing domain adaptation and catastrophic forgetting in adverse driving conditions.

Findings

PSS effectively adapts to multiple adverse conditions without forgetting previous domains.

The approach generalizes well to unseen adverse conditions.

Extensive evaluations show improved segmentation performance across datasets.

Abstract

Semantic segmentation for autonomous driving is an even more challenging task when faced with adverse driving conditions. Standard models trained on data recorded under ideal conditions show a deteriorated performance in unfavorable weather or illumination conditions. Fine-tuning on the new task or condition would lead to overwriting the previously learned information resulting in catastrophic forgetting. Adapting to the new conditions through traditional domain adaption methods improves the performance on the target domain at the expense of the source domain. Addressing these issues, we propose an architecture-based domain-incremental learning approach called Progressive Semantic Segmentation (PSS). PSS is a task-agnostic, dynamically growing collection of domain-specific segmentation models. The task of inferring the domain and subsequently selecting the appropriate module for segmentation is carried out using a collection of convolutional autoencoders. We extensively evaluate our proposed approach using several datasets at varying levels of granularity in the categorization of adverse driving conditions. Furthermore, we demonstrate the generalization of the proposed approach to similar and unseen domains.