MSeg

该数据集名为MSeg，是一个强大的分割网络，它为合成图像和真实图像提供语义分割标签。此外，该数据集被用于生成用于训练和评估所提出的光照真实感增强方法的语义地图。其任务是进行语义分割，以提升图像的真实感。

This dataset, named MSeg, is a powerful segmentation network that provides semantic segmentation labels for both synthetic and real-world images. Additionally, this dataset is used to generate semantic maps for training and evaluating the proposed illumination realism enhancement method. Its task is semantic segmentation to improve the realism of images.