CrisisMMD
收藏资源简介:
CrisisMMD是由卡塔尔计算研究机构创建的大型多模态数据集,包含约1422万条推文和57.6万张图片,数据来源于2017年全球七次重大自然灾害。数据集通过特定的关键词和标签从Twitter收集,经过人工标注,用于支持人道主义援助中的多种任务,如信息分类、损害评估等。该数据集旨在通过提供丰富的文本和图像数据,帮助研究人员开发更有效的灾害响应和管理系统。
CrisisMMD is a large-scale multimodal dataset created by the Qatar Computing Research Institute. It encompasses roughly 14.22 million tweets and 576,000 images, with data sourced from seven major global natural disasters occurring in 2017. The dataset was collected from Twitter using specific keywords and tags, and underwent manual annotation. It is designed to support diverse tasks in humanitarian assistance, such as information classification and damage assessment. The primary objective of this dataset is to provide abundant textual and visual data to aid researchers in developing more effective disaster response and management systems.
- 1CrisisMMD: Multimodal Twitter Datasets from Natural Disasters卡塔尔计算研究机构,HBKU,多哈 · 2018年



