Domain Adaptation with Morphologic Segmentation

TL;DR

This paper introduces a domain adaptation method using morphologic segmentation to convert diverse images into a uniform, realistic domain, improving training data quality for computer vision tasks across real and synthetic urban scene datasets.

Contribution

The paper proposes a novel framework that transforms images into a generalized EPS representation before converting them into a unified output domain, enhancing data consistency for training.

Findings

Images are photo-realistic and artifact-free after transformation.

The method improves training data diversity and reduces overfitting.

Effective on multiple urban scene datasets, both real and synthetic.

Abstract

We present a novel domain adaptation framework that uses morphologic segmentation to translate images from arbitrary input domains (real and synthetic) into a uniform output domain. Our framework is based on an established image-to-image translation pipeline that allows us to first transform the input image into a generalized representation that encodes morphology and semantics - the edge-plus-segmentation map (EPS) - which is then transformed into an output domain. Images transformed into the output domain are photo-realistic and free of artifacts that are commonly present across different real (e.g. lens flare, motion blur, etc.) and synthetic (e.g. unrealistic textures, simplified geometry, etc.) data sets. Our goal is to establish a preprocessing step that unifies data from multiple sources into a common representation that facilitates training downstream tasks in computer vision. This way, neural networks for existing tasks can be trained on a larger variety of training data, while they are also less affected by overfitting to specific data sets. We showcase the effectiveness of our approach by qualitatively and quantitatively evaluating our method on four data sets of simulated and real data of urban scenes. Additional results can be found on the project website available at http://jonathank.de/research/eps/ .