海天瑞声-人像姿态表情视频数据库

数据集包含口播数字人、唇动数据集，广泛应用于数字人、虚拟主播、在线教育等场景。涵盖了视频、图像、音频和文本等多种数据类型，并经过高质量采集和精细标注，确保数据的准确性和一致性。

This dataset comprises two subsets: the talking digital human dataset and the lip movement dataset. It is widely used in scenarios such as digital humans, virtual anchors, and online education. The dataset covers multiple data types including video, image, audio, and text, and has been collected with high quality and meticulously annotated to ensure the accuracy and consistency of the data.

搜集汇总

背景与挑战

背景概述

该数据集是一个关于人像姿态表情的视频数据库，采用Apache License 2.0许可，可通过ModelScope SDK或GIT Clone命令下载，但详情页面未提供具体数据内容和特点的详细描述。