UIEB

该数据集由UIEB提供，包含800对训练数据和90对测试数据，专门用于评估水下图像增强算法的性能。此外，该数据集还用于比较HIFI-Net网络与现有先进算法的表现。在规模上，训练集包含800对数据，而测试集包含90对数据。该数据集的任务是专注于水下图像增强。

This dataset, provided by UIEB, consists of 800 pairs of training data and 90 pairs of test data, and is specifically dedicated to evaluating the performance of underwater image enhancement algorithms. Additionally, this dataset is employed to compare the performance of the HIFI-Net network against existing state-of-the-art algorithms. Regarding its scale, the training set includes 800 pairs of data, whereas the test set contains 90 pairs. The core task of this dataset focuses on underwater image enhancement.

搜集汇总

背景与挑战

背景概述

UIEB是一个包含950张真实水下图像的基准数据集，其中890张有参考图像，60张为挑战性图像，用于评估水下图像增强算法，并提供基线模型Water-Net。