---
license: apache-2.0
datasets:
- OpenIllumination/OpenIllumination
language:
- en
- ko
pipeline_tag: image-to-image
---
# 🔦LightTransporter🚀​ (2024 Fall CS492D Final Project)  

[**Poster**](https://drive.google.com/file/d/101eOfUqfS-wKGuJPvaqoBtKvs4GrIywM/view?usp=sharing) | [**Report**](https://drive.google.com/file/d/1009QskpqLxJRltqiCnim5seDmX1nxwEL/view?usp=sharing) | [**Github**](https://github.com/j-mayo/LightTransporter/)
> <b>KAIST 2024 Fall CS492D: Diffusion Models and Their Applications</b>

![Project_Poster_20170693_20244300_20244496](https://github.com/user-attachments/assets/5a5c0e45-d6f8-4942-91e5-7245f3c0f5f7)

## Introduction
This is a **Team03** official repository of the project in the KAIST 2024 Fall CS492D lecture, [Diffusion Models and Their Applications](https://mhsung.github.io/kaist-cs492d-fall-2024/).

## Dataset
We used [OpenIllumination](https://oppo-us-research.github.io/OpenIllumination/) dataset for our project, and [LiT](https://github.com/KAIST-Visual-AI-Group/Diffusion-Project-Illumination) project repo for downloading and pre-processing. 

The image below shows the environment map of each light condition, which we used to encode the light condition.

<img src="https://github.com/user-attachments/assets/7e51df4c-8585-4167-8db6-78e9ecbf1032" width="500" height="200">

## Qualitative Results
Each row shows the **source**, **target**(ground truth), and **generated images** from left to right.

![image](https://github.com/user-attachments/assets/33972a3c-086b-4fbd-8df9-f09f75930580)


## Quantitative Results
![image](https://github.com/user-attachments/assets/a1d5c9e3-9097-4cf2-b418-4de08e17bbdf)

## Team Information
- [Junwoo Choi](https://github.com/Str4Strength) (School of Computing, KAIST)
- [Jaeyo Shin](https://github.com/j-mayo) (Graduate School of AI, KAIST)
- [Dongwon Choi](https://github.com/chlehdwon) (Graduate School of AI, KAIST)