Learning photographic global tonal adjustment with a database of input / output image pairs

Following the setup of previous works [8, 16], we conducted experiments on various bit image restoration tasks.We utilized a dataset of 2000 16-bit images, with trainingdata sourced from SINTEL [37] and FIVE-K [38]. SINTELis an animated short film dataset containing over 20,000 16-bit lossless images with a resolution of 436 × 1024 pixels. InFIVE-K, randomly select images from 5,000 16-bit naturalimages for the experiment.The test set includes 8 imagesrandomly chosen from the SINTEL dataset (referred to asSINTEL-8） 和 40 张从 FIVEK 数据集（称为 FIVEK-40）、（FIVEK-40 和SINTEL-8 数据集是独立的公共数据集的训练集。）与训练集没有重叠。