Development and Evaluation of a Novel Marker-Based Tracking System for 3D-Scaled Masonry Models using DeepTag

This paper presents the development and evaluation of a novel marker-based tracking system leveraging deep learning techniques, i.e., DeepTag, for the monitoring of 3D-scaled masonry models. DeepTag represents a significant advancement over previous methodologies by reducing measurement uncertainty by 42% and enhancing tracking accuracy with fewer markers. Utilizing convolutional neural networks (CNNs), DeepTag effectively addresses issues such as occlusions and varying lighting conditions, making it suitable for structural monitoring. The system's robust performance in tracking deformations and displacements highlights its potential for the preservation and maintenance of masonry structures. Preliminary experimental results underscore the system's viability as a low-cost and efficient alternative to traditional tracking systems.