Table 1 Research on using deep learning algorithms to detect different cultural heritage damages
From: Using mask R-CNN to rapidly detect the gold foil shedding of stone cultural heritage in images
Researchers | Algorithm | Research object | Specific damages |
|---|---|---|---|
Chaiyasarn et al. [24] | CNN, SVM | Historical masonry structures of a stupa in Wat Chai Watthanaram, Thailand | Crack |
Kwon et al. [25] | Faster R-CNN | Outdoor stone cultural properties conducted by the Cultural Heritage Administration of S. Korea | Crack, loss, detachment, biological colonization |
Wang et al. [26] | Faster R-CNN, Mask R-CNN | Historic glazed tiles of the Palace Museum in China | Surface damage |
Wang et al. [34] | Faster R-CNN | Historic brick masonry structures of the Palace Museum | Efflorescence, spalling |
Zou et al. [27] | Faster R-CNN | Components of the Forbidden City | The missing parts (Goutou, Dishui, Dingmao) |
ANGHELUŢĂ et al. [36] | VGG-16 specialized convolutional network | A wood painting representing St. Constantine and Helen | Crack, blister, detachment |
Hatır et al. [29] | Mask R-CNN | Yazılıkaya monuments in the Hattusa archeological site | Biological colonization, contour scaling, crack, higher plant, impact damage, microkarst, missing part |
Adamopoulos et al. [23] | Supervised segmentation methods based on random decision trees, ensemble learning, and regression algorithm | A fortification in Euboea, Greece | Vegetation, moss, black crusts, lichens, missing material, dampness |