概述
视频-文本合规(Video-Text Compliance,VTC) 包含 7920 个样本,每个样本由一对视频-文本说明和一个合规/不合规标签组成。此数据集包含了超过 120 万帧数据。我们在数据收集方面采取了一种独特的方法,可以通过一组核心视频自动扩充数据集。为了回应人们对数据隐私日益增长的关注度,我们在生成 VTC 数据集时认真遵循了隐私保护安全措施。
数据集元数据
记录示例
carry_bag_P1000344_iter006.mp4 0 open_predetermined_suitcase_calmly
carry_bag_P1000344_iter007.mp4 0 precisely_place_the_appropriate_box
carry_bag_P1000344_iter005.mp4 0 push_accessible_cart
carry_bag_P1000344_iter004.mp4 0 open_the_applicable_bag_at_once
carry_bag_P1000344_iter000.mp4 0 carry_the_specified_box
引用
@InProceedings{Jaiswal_2019_ICCV_Workshops,
author = {Jaiswal, Mayoore and Liu, Frank and Jagannathan, Anupama and Gattiker, Anne and Hwang, Inseok and Lee, Jinho and Tong, Matthew and Dureja, Sahil and Shah, Soham and Hofstee, Peter and Chen, Valerie and Paul, Suvadip and Feris, Rogerio},
title = {Video-Text Compliance: Activity Verification Based on Natural Language Instructions},
booktitle = {The IEEE International Conference on Computer Vision (ICCV) Workshops},
month = {Oct},
year = {2019}
}
相关链接
- Video-Text Compliance: Activity Verification Based on Natural Language Instructions(论文) Video-Text Compliance (VTC) 数据集包含原子活动的视频,以及文本说明和合规标签。VTC 数据集通过自动扩充技术而构建,可保护隐私,包含超过 120 万帧数据。
本文翻译自:Video-Text Compliance(2019-10-24)