BVI-Artefact

BVI-Artefact是由布里斯托大学视觉信息实验室创建的一个大型数据库，专门用于检测流媒体视频中的视觉伪影。该数据集包含480个视频序列，每个序列包含多种伪影，并附有二进制伪影标签。数据集的创建旨在模拟真实世界的视频流媒体情况，包括采集、后期制作、压缩和传输等不同阶段的伪影。BVI-Artefact的应用领域主要集中在优化流媒体用户体验，通过识别和定位伪影来控制流媒体质量。BVI-Artefact is a large-scale database developed by the Visual Information Laboratory at the University of Bristol, specifically dedicated to detecting visual artifacts in streaming video. This dataset comprises 480 video sequences, each containing multiple types of visual artifacts and accompanied by binary artifact labels. The dataset was created to simulate real-world video streaming scenarios, covering artifacts generated during different stages such as acquisition, post-production, compression, and transmission. The primary applications of BVI-Artefact focus on optimizing streaming user experience by identifying and localizing visual artifacts to control streaming quality.

搜集汇总

背景与挑战

背景概述

BVI-Artefact是由布里斯托大学视觉信息实验室创建的大型数据库，包含480个视频序列，每个序列带有多种伪影和二进制标签，旨在模拟真实世界流媒体视频的采集、后期制作、压缩和传输阶段的伪影。该数据集主要用于优化流媒体用户体验，通过识别和定位视觉伪影来控制流媒体质量。