ResNet50V2 for Global Landmark Categorization: A Deep Transfer Learning and Residual Mapping Approach

A deep learning model, ResNet50V2, demonstrates high accuracy in classifying 12 globally significant landmarks, including the Historic Sanctuary of Machu Picchu. This advancement leverages transfer learning and residual mapping to handle complex variations in form, history, and location, addressing challenges posed by diverse cultural heritage sites. The model's ability to generalize and extract relevant features supports applications in tourism marketing, heritage conservation, digital preservation, and educational resources. By automating landmark recognition, the approach promotes sustainable management of cultural heritage and enhances global learning initiatives.

Dey, R., Banerjee, D., Sharma, J., Singla, M. K., Singh, A. K., Chitre, A., & Mahajan, R. A. (2025). ResNet50V2 for Global Landmark Categorization: A Deep Transfer Learning and Residual Mapping Approach. 2025 IEEE 5th International Conference on ICT in Business Industry &Amp; Government (ICTBIG), 1–5. https://doi.org/10.1109/ictbig68706.2025.11323705