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

Official PyTorch repository for StawGAN: Structural-Aware Generative Adversarial Networks for Infrared Image Translation, ISCAS 2023

[IEEEXplore][ArXiv preprint]

Luigi Sigillo, Eleonora Grassucci, and Danilo Comminiello

The link to the website presentation is BlogPost

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.

Model Architecture 🎬

How to run experiments 💻

First, please install the requirements:

pip install -r requirements.txt

Cite

Please cite our work if you found it useful:

@INPROCEEDINGS{10181838,
  author={Sigillo, Luigi and Grassucci, Eleonora and Comminiello, Danilo},
  booktitle={2023 IEEE International Symposium on Circuits and Systems (ISCAS)}, 
  title={StawGAN: Structural-Aware Generative Adversarial Networks for Infrared Image Translation}, 
  year={2023},
  volume={},
  number={},
  pages={1-5},
  doi={10.1109/ISCAS46773.2023.10181838}}