S2R-HDR
收藏资源简介:
S2R-HDR是一个由上海人工智能实验室和香港中文大学合作创建的大型高质仿真HDR图像数据集,包含24000个样本。该数据集使用虚幻引擎5设计了多样化的真实HDR场景,涵盖各种动态元素、运动类型和高动态范围场景。数据集在创建过程中,通过自定义渲染管线,确保输出数据保持在线性HDR空间,并以EXR格式存储,以防止数据量化。S2R-HDR数据集旨在用于HDR融合,并解决动态场景下HDR重建的泛化问题。
S2R-HDR is a large-scale high-quality simulated HDR image dataset co-developed by Shanghai AI Laboratory and The Chinese University of Hong Kong, which contains 24,000 samples. This dataset uses Unreal Engine 5 to design diverse realistic HDR scenes, covering various dynamic elements, motion types and high dynamic range scenarios. During the dataset creation process, a custom rendering pipeline was adopted to ensure that the output data remains in the linear HDR space, and is stored in EXR format to avoid data quantization. The S2R-HDR dataset is intended for HDR fusion tasks, and aims to address the generalization problem of HDR reconstruction in dynamic scenes.




