GAN-Based Tone Curve Learning Approach to HDR Image Reconstruction Method Using Single LDR Image
In recent years, the use of HDR images is becoming more widespread. In this study, we discuss a method for converting an already taken LDR image into an HDR image. HDR generating methods based on the deep learning have been well studied in recent years. However, these methods have problems such as adding non-existent information on the image. In this paper, using single LDR image and a few reference LDR images, we propose a new HDR generation method based on a GAN to calculate the tone curves. The advantage of the proposed method is that there are no restrictions on the size or scene of the input and reference images, and they can differ. Numerical experiments show the effectiveness of the proposed method.
Authors
- Kazunori Uruma (ORCID: https://orcid.org/0000-0002-3915-6796)
- Ryohei Sasaki (ORCID: https://orcid.org/0000-0003-2327-7910)
- Tomoya Yoshida (ORCID: https://orcid.org/0000-0002-9078-0252)
Institutions
- Kogakuin University (JP)
- Tokyo University of Technology (JP)
Publication Details
- Published
- 2024-12-01
- DOI
- https://doi.org/10.1109/tencon61640.2024.10902822
- Primary Topic
- Image Enhancement Techniques
- Type
- article
- Field-Weighted Citation Impact
- 0.00