A Real-world Display Inverse Rendering Dataset
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908495676178432 |
|---|---|
| author | Choi, Seokjun Chung, Hoon-Gyu Jeon, Yujin Nam, Giljoo Baek, Seung-Hwan |
| author_facet | Choi, Seokjun Chung, Hoon-Gyu Jeon, Yujin Nam, Giljoo Baek, Seung-Hwan |
| contents | Inverse rendering aims to reconstruct geometry and reflectance from captured images. Display-camera imaging systems offer unique advantages for this task: each pixel can easily function as a programmable point light source, and the polarized light emitted by LCD displays facilitates diffuse-specular separation. Despite these benefits, there is currently no public real-world dataset captured using display-camera systems, unlike other setups such as light stages. This absence hinders the development and evaluation of display-based inverse rendering methods. In this paper, we introduce the first real-world dataset for display-based inverse rendering. To achieve this, we construct and calibrate an imaging system comprising an LCD display and stereo polarization cameras. We then capture a diverse set of objects with diverse geometry and reflectance under one-light-at-a-time (OLAT) display patterns. We also provide high-quality ground-truth geometry. Our dataset enables the synthesis of captured images under arbitrary display patterns and different noise levels. Using this dataset, we evaluate the performance of existing photometric stereo and inverse rendering methods, and provide a simple, yet effective baseline for display inverse rendering, outperforming state-of-the-art inverse rendering methods. Code and dataset are available on our project page at https://michaelcsj.github.io/DIR/ |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2508_14411 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | A Real-world Display Inverse Rendering Dataset Choi, Seokjun Chung, Hoon-Gyu Jeon, Yujin Nam, Giljoo Baek, Seung-Hwan Graphics Computer Vision and Pattern Recognition Inverse rendering aims to reconstruct geometry and reflectance from captured images. Display-camera imaging systems offer unique advantages for this task: each pixel can easily function as a programmable point light source, and the polarized light emitted by LCD displays facilitates diffuse-specular separation. Despite these benefits, there is currently no public real-world dataset captured using display-camera systems, unlike other setups such as light stages. This absence hinders the development and evaluation of display-based inverse rendering methods. In this paper, we introduce the first real-world dataset for display-based inverse rendering. To achieve this, we construct and calibrate an imaging system comprising an LCD display and stereo polarization cameras. We then capture a diverse set of objects with diverse geometry and reflectance under one-light-at-a-time (OLAT) display patterns. We also provide high-quality ground-truth geometry. Our dataset enables the synthesis of captured images under arbitrary display patterns and different noise levels. Using this dataset, we evaluate the performance of existing photometric stereo and inverse rendering methods, and provide a simple, yet effective baseline for display inverse rendering, outperforming state-of-the-art inverse rendering methods. Code and dataset are available on our project page at https://michaelcsj.github.io/DIR/ |
| title | A Real-world Display Inverse Rendering Dataset |
| topic | Graphics Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2508.14411 |