Automatic analysis of social media images to identify disaster type and infer appropriate emergency response

Journal of Big Data

Sherif Abdou¹  Khaled Mostafa Elsayed²  Mohsen Rashwan³  Amna Asif⁴  Md Maruf Hasan⁴  Majed A. Alshamari⁴  Shaheen Khatoon⁴ 

[1] Faculty of Computers and Artificial Intelligence, Cairo University, Giza, Egypt;Faculty of Computing & Information, Cairo University, Giza, Egypt;Faculty of Engineering, Cairo University, Giza, Egypt;Information Systems Department, College of Computer Sciences & Information Technology, King Faisal University (KFU), AlAhsa, Kingdom of Saudi Arabia;

关键词: Convolutional neural networks;    Image classification;    Object detection;    Disaster management;   

DOI  :  10.1186/s40537-021-00471-5

来源: Springer

【 摘 要 】

Social media postings are increasingly being used in modern days disaster management. Along with the textual information, the contexts and cues inherent in the images posted on social media play an important role in identifying appropriate emergency responses to a particular disaster. In this paper, we proposed a disaster taxonomy of emergency response and used the same taxonomy with an emergency response pipeline together with deep-learning-based image classification and object identification algorithms to automate the emergency response decision-making process. We used the card sorting method to validate the completeness and correctness of the disaster taxonomy. We also used VGG-16 and You Only Look Once (YOLO) algorithms to analyze disaster-related images and identify disaster types and relevant cues (such as objects that appeared in those images). Furthermore, using decision tables and applied analytic hierarchy processes (AHP), we aligned the intermediate outputs to map a disaster-related image into the disaster taxonomy and determine an appropriate type of emergency response for a given disaster. The proposed approach has been validated using Earthquake, Hurricane, and Typhoon as use cases. The results show that 96% of images were categorized correctly on disaster taxonomy using YOLOv4. The accuracy can be further improved using an incremental training approach. Due to the use of cloud-based deep learning algorithms in image analysis, our approach can potentially be useful to real-time crisis management. The algorithms along with the proposed emergency response pipeline can be further enhanced with other spatiotemporal features extracted from multimedia information posted on social media.

【 授权许可】

CC BY