MLS (Multiple Light Source)

用于颜色研究的多光源数据集
24 个多物体场景的图像。
每个场景都是在 18 个多光源照明场景下拍摄的，光源的主要光谱颜色、强度和与场景的距离各不相同。
相机的光谱特性、光源和均匀着色的物体表面。
每个场景的颜色均匀的对象表面的逐像素蒙版。
我们主要解决计算颜色恒常性算法评估的现实场景，但也旨在使数据尽可能通用，用于计算颜色科学和计算机视觉。详细信息已发表在：用于颜色研究的多光源数据集。

Multi-light Source Dataset for Color Research
Images of 24 multi-object scenes.
Each scene was captured under 18 multi-light illumination conditions, with the dominant spectral color, intensity, and distance of the light sources to the scene varying.
Spectral characteristics of the camera, light sources, and uniformly colored object surfaces.
Per-pixel masks of the color-uniform object surfaces for each scene.
This dataset is primarily designed for evaluating computational color constancy algorithms in real-world scenarios, while also aiming to make the data as generalizable as possible for use in computational color science and computer vision. Detailed information has been published in: Multi-light Source Dataset for Color Research.

搜集汇总

背景与挑战

背景概述

MLS (Multiple Light Source) 是一个用于颜色研究的多光源数据集，包含24个多物体场景图像，每个场景在18种不同光源条件下拍摄，主要应用于计算颜色恒常性算法评估和计算机视觉领域。该数据集由俄罗斯科学院于2019年发布，采用CC BY-SA 4.0授权协议。