Tencent Ads Video Segmentation (TAVS)
收藏arXiv2022-12-09 更新2024-08-06 收录
下载链接：
http://arxiv.org/abs/2212.04700v1
下载链接
链接失效反馈官方服务：
资源简介：
TAVS数据集是由腾讯数据平台创建，专注于广告视频的多模态场景分割。该数据集包含12,000个视频，覆盖教育、电子商务、视频游戏等多个行业领域，旨在通过视频、音频和文本等多模态信息，实现视频内容的全面理解和精细分割。TAVS数据集通过三层类别体系，从‘展示形式’、‘地点’和‘风格’三个独立视角描述视频，支持多标签分类，适用于视频编辑、推荐系统等实际应用场景。
The TAVS dataset was created by Tencent Data Platform, focusing on multimodal scene segmentation for advertising videos. It includes 12,000 videos covering multiple industry sectors such as education, e-commerce, video games and other fields. The dataset aims to achieve comprehensive understanding and precise segmentation of video content via multimodal information including video, audio and text. With a three-layer category system, the TAVS dataset describes videos from three independent perspectives: presentation format, setting and style, and supports multi-label classification, which is applicable to practical application scenarios like video editing and recommendation systems.
提供机构：
腾讯数据平台
创建时间：
2022-12-09
