Incidents1M Dataset
收藏资源简介:
Incidents1M数据集是由麻省理工学院计算机科学与人工智能实验室创建的大型多标签数据集,包含977,088张图像,涉及43种事故和49个地点类别。数据集通过从Google图像查询中下载图像,并使用Amazon Mechanical Turk进行手动标注,以确保图像与特定事故或地点标签相关。该数据集旨在通过自动化图像过滤技术,提高对自然灾害发生后情况的快速理解和响应,特别是在社交媒体图像分析方面。应用领域包括灾害分析、事故检测和实时监控,以支持人道主义援助和灾害响应。
Incidents1M is a large-scale multi-label dataset developed by the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL). It contains 977,088 images, covering 43 accident categories and 49 location categories. The dataset is constructed by downloading images from Google Image searches and conducting manual annotations via Amazon Mechanical Turk, to ensure that each image is relevant to a specific accident or location label. This dataset aims to improve rapid understanding and response to post-disaster situations through automated image filtering technologies, with a particular focus on social media image analysis. Its application areas include disaster analysis, accident detection, and real-time monitoring, to support humanitarian aid and disaster response work.




