LOL-v2

该数据集被划分为真实和合成两个子集，分别用于训练和测试。它包含了详尽的训练和测试配置信息。在规模上，该数据集提供了689张真实世界的图像用于训练，以及100张用于测试。所涉及的任务是针对低光照场景的图像增强。

This dataset is divided into two subsets: real-world and synthetic, which are respectively used for training and testing. It contains comprehensive configuration information for both training and testing stages. In terms of scale, the dataset provides 689 real-world images for training and 100 images for testing. The task involved is image enhancement for low-light scenarios.

搜集汇总

背景与挑战

背景概述

LOL-v2是一个专注于低光图像增强的数据集，包含真实和合成的低光与正常光图像，支持半监督学习方法，旨在通过深度递归带网络（DRBN）技术提升图像的视觉质量。