Table 1.

Examples of Transfer Learning (TL) Applications in SHM using pre-trained deep neural networks (DNNs).

Pre-Trained Network	Purpose/Application	Researches in SHM
VGG (VGG-16, 19) [17]	Crack detection Mixed reality systems Bolt loosening detection Corrosion detection Component recognition Steel damage condition assessment Post-earthquake assessment	[32,34,102,141,144,145,176,178,186,187,188,189,190,191,192,193,194]
Inception (Inception-V2, V3, V4) 1 [19,181,195]	Crack detection Damage detection of historic masonry buildings	[33,169,176,196,197]
ResNet (ResNet-20, 50, 101, 152) [185]	Crack detection Bridge component extraction Structural inspection	[33,139,144,171,178,190,191,194,198,199,200]
AlexNet [18]	Crack detection Comprehensive maintenance and inspection sUAS-assisted structural inspections Post-earthquake assessment	[73,96,122,124,143,178,193,196,201,202,203]
GoogleNet [19]	Crack detention Post-disaster inspection Comprehensive maintenance and inspection for bridges	[143,157,169,178,196,198,202]
MobileNet	Road damage detection	[173]
UNet, SegCaps, SegNet [113]	Segmentation Pixel-level crack detection	[186,194]
ZF-net [187]	ZF-Net for fast R-CNN. Region-based DL for detecting multiple damage types Detection and localization of multiple types of damage Comprehensive maintenance and inspection for bridges Volumetric damage quantification	[167,203,204,205]
CrackNet, CrackNet-R [139,188]	Crack detection Pixel-level road crack detection	[27,101,194]

¹ V denotes the version.