| --- |
| license: other |
| license_name: pexels |
| license_link: https://www.pexels.com/license/ |
| --- |
| |
| **🤗We will publicly release the dataset upon paper acceptance.🤗** |
|
|
| # TriLayer dataset for layered video generation training |
| Github: [DBL-Diffusion](https://github.com/KyujinHan/DBL-Diffusion) |
| Paper: [DBL-Diffusion](https://arxiv.org/pdf/2607.25802) |
|
|
|
|
| # Introduction |
|  |
| To support learning layered representations that capture both object appearance and object-induced visual effects, TriLayer provides aligned composite, background, and foreground videos for each sample. |
| The composite video contains the original scene with the object present. |
| The foreground video and its alpha matte capture both opaque object regions and semi-transparent effects such as shadows and reflections. |
| The background video contains neither the object nor its associated effects, serving as a clean reference for decomposition and as the input for layered object insertion. |
| Although each sample contains three aligned videos, these are not independent layers; the composite is physically formed by alpha-compositing the foreground layer onto the background. |
| Each sample additionally provides the object name and a VLM-generated caption describing its appearance and associated effects, which serve as conditioning signals for both DBL-Insert and DBL-Decompose. |
| The dataset contains **3,964 video triplets** spanning diverse objects, motions, environments, and lighting conditions. |
|
|
| # Dataset Info |
| (TODO) |
|
|
| # License |
|
|
| All data is licensed by [Pexels](https://www.pexels.com/license/). |
| To best our knowledge, the Pexels dataset may be used for commercial purposes. |
| Please read this [license page](https://help.pexels.com/hc/en-us/articles/360042295174-What-is-the-license-of-the-photos-and-videos-on-Pexels) and double-check. |
