Assessment of indoor risk through deep learning -based object recognition in disaster situations

Irshad Khan, Ziyi Guo, Kihwan Lim, Jaeseon Kim, Young Woo Kwon

Research output: Contribution to journalArticlepeer-review

Abstract

Disasters can devastate individuals and their properties, highlighting the importance of risk assessment to promote safety. Recently, deep learning techniques have shown the potential in identifying hazardous situations during disasters. Recognizing potentially dangerous objects in indoor environments can be essential for assisting individuals in responding appropriately to emergencies. In this article, we present an indoor-risk analysis framework for disasters based on deep learning. Our framework utilizes modern deep learning techniques to calculate an indoor risk rating based on dangerous objects’ sizes, enabling comprehensive risk assessment of indoor environments during disasters. To that end, we use (Mask R-CNN) to identify hazardous indoor objects in disaster situations with 94% accuracy. By incorporating object size information, our framework offers a more nuanced and detailed risk assessment than previous approaches. Our proposed system provides a valuable tool for promoting ongoing safety improvement and enhancing indoor safety during natural disasters.

Original languageEnglish
Pages (from-to)34669-34690
Number of pages22
JournalMultimedia Tools and Applications
Volume83
Issue number12
DOIs
StatePublished - Apr 2024

Keywords

  • Disaster assessment
  • Indoor safety
  • Object recognition
  • Risk level

Fingerprint

Dive into the research topics of 'Assessment of indoor risk through deep learning -based object recognition in disaster situations'. Together they form a unique fingerprint.

Cite this