採択課題 【詳細】
| jh250017 | 次世代災害対応のための視覚言語モデルの構築 |
|---|---|
| 課題代表者 | 横矢直人(東京大学大学院新領域創成科学研究科) Naoto Yokoya (Graduate School of Frontier Sciences, The University of Tokyo) |
| 概要 | Sudden-onset disasters occur frequently worldwide, posing a serious threat to human life and property safety. In the last few decades, natural disasters, i.e., earthquakes and storms to floods and droughts kill approximately 40,000 to 50,000 people per year. For effective humanitarian assistance and disaster response, we plan to leverage earth observation and artificial intelligence techniques, developing cutting-edge disaster vision-language models for macro-level damage assessment and micro-level analysis of disaster-bearing bodies. We believe our developed vision-language models as well as constructed multi-modal datasets could provide a solid benchmark that promotes rapid disaster response. |
| 報告書等 | 研究紹介ポスター ／ 最終報告書 |
| 関連Webページ |
無断転載禁止