mos分评测数据集

### Clone with HTTP ```bash git clone https://www.modelscope.cn/datasets/charlesHuang/vqa_mos_youtubeUGC_validation.git ``` ## YOUTUBE UGC 图像质量评价数据集 ### 数据集描述 YOUTUBE UGC 图像质量评价数据集包含多个类别场景，多种自然噪声，多种编解码瑕疵，拍摄者引入的曝光、运动模糊等瑕疵，以及对应的mos分。 当前数据集为源数据集的子集，并通过抽帧、筛选得到的图像序列和对应的MOS分。 ### 原始地址 > https://media.withyoutube.com/bash git clone https://www.modelscope.cn/datasets/charlesHuang/vqa_mos_youtubeUGC_validation.git ## YouTube UGC Image Quality Assessment Dataset ### Dataset Description The YouTube UGC Image Quality Assessment Dataset encompasses multiple scene categories, various natural noises, diverse coding and decoding artifacts, artifacts introduced by content creators such as exposure defects and motion blur, as well as corresponding Mean Opinion Score (MOS) values. This dataset is a subset of the original source dataset, consisting of image sequences and their matching MOS scores obtained through frame extraction and screening. ### Original Source Address > https://media.withyoutube.com/

搜集汇总

背景与挑战

背景概述

该数据集为YOUTUBE UGC图像质量评估数据集，包含多种场景、自然噪声、编解码缺陷和拍摄瑕疵（如曝光、运动模糊等），并提供了对应的MOS分数。作为原始数据集的子集，它通过帧提取和筛选获得图像序列及其MOS评分。