 This study presents a multimodal integrated structure that combines ortho-photo and off-nadi images for earthquake-building damage detection using deep transfer learning, achieving an overall accuracy of over 83%. This article was authored by Gaysam Abdi and Shabnam Jabari.