# Don't Worry About the Weather: Unsupervised Condition-Dependent Domain Adaptation

## TL;DR

This paper introduces a lightweight, unsupervised domain adaptation system that preprocesses images to maintain high performance of existing vision models across varying conditions without fine-tuning.

## Contribution

It proposes a novel input adapter approach that enables off-the-shelf models to adapt to different conditions without retraining, using self-supervised incremental learning.

## Key findings

- Significant improvements in semantic segmentation accuracy.
- Enhanced topological localization performance.
- Effective adaptation across diverse weather and lighting conditions.

## Abstract

Modern models that perform system-critical tasks such as segmentation and localization exhibit good performance and robustness under ideal conditions (i.e. daytime, overcast) but performance degrades quickly and often catastrophically when input conditions change. In this work, we present a domain adaptation system that uses light-weight input adapters to pre-processes input images, irrespective of their appearance, in a way that makes them compatible with off-the-shelf computer vision tasks that are trained only on inputs with ideal conditions. No fine-tuning is performed on the off-the-shelf models, and the system is capable of incrementally training new input adapters in a self-supervised fashion, using the computer vision tasks as supervisors, when the input domain differs significantly from previously seen domains. We report large improvements in semantic segmentation and topological localization performance on two popular datasets, RobotCar and BDD.