StawGAN: Structural-Aware Generative Adversarial Networks for Infrared Image Translation

TL;DR

StawGAN is a novel structural-aware GAN that improves infrared to daytime image translation by enhancing object shape and detail, outperforming existing models on aerial datasets.

Contribution

The paper introduces StawGAN, a new model that incorporates structural awareness to produce higher quality infrared to daytime image translations.

Findings

Outperforms state-of-the-art image translation models on DroneVeichle dataset.

Produces more accurate and high-definition object shapes in translated images.

Effectively enhances the quality of infrared to daytime image translation.

Abstract

This paper addresses the problem of translating night-time thermal infrared images, which are the most adopted image modalities to analyze night-time scenes, to daytime color images (NTIT2DC), which provide better perceptions of objects. We introduce a novel model that focuses on enhancing the quality of the target generation without merely colorizing it. The proposed structural aware (StawGAN) enables the translation of better-shaped and high-definition objects in the target domain. We test our model on aerial images of the DroneVeichle dataset containing RGB-IR paired images. The proposed approach produces a more accurate translation with respect to other state-of-the-art image translation models. The source code is available at https://github.com/LuigiSigillo/StawGAN