MSNet: A Multilevel Instance Segmentation Network for Natural Disaster Damage Assessment in Aerial Videos

Video | Hurricane and Tornado | Response

ISBDA (Instance Segmentation in Building Damage Assessment) is a dataset for hurricane and tornado building damage assessment. It contains 1,030 images from 10 videos of disaster aftermaths (84 min total duration) and 2,961 damaged part instances. It is introduced for image segmentation and multiclass (ordinal) classification.

  • ML task type: Image segmentation, multiclass (ordinal) classification
  • Data Source: Social Media
  • Size: 1,030 Images sampled from 10 videos (84 min total duration)
  • Timespan: 2017 - 2019
  • Geographical Coverage: Florida, Missouri, Illinois, Texas, Alabama, North Carolina
  • Baseline Information
  • Evaluated on: MS-Net
  • Metrics used: COCO instance segmentation metric including AP (averaged over all IoU thresholds)
  • Results as reported in original paper: 37.2

Xiaoyu Zhu, Junwei Liang, and Alexander Hauptmann. MSNET: A multilevel instance segmentation network for natural disaster damage assessment in aerial videos. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pages 2023–2032, 2021.