Alexander Tzitamidis

An incident recording application was developed using Augmented Reality (AR) technology to digitally preserve environments like accident scenes. The system was built on the Unity3D engine, where Project Tango and SLAM are utilized for real-time spatial mapping. Immersive 3D models are created through motion tracking and depth perception, enabling detailed investigations without being physically present.

A mobile application for cultural heritage preservation is presented, leveraging SLAM, depth mapping, and 360-degree imaging. Developed via Unity, C#, and Firebase, the system captures high-fidelity spatial data and 3D reconstructions. It provides curators with annotated virtual environments to analyze post-disaster scene dynamics and assist in the restoration of damaged artifacts.

A specialized Object Detection dataset is presented and hosted on Zenodo, featuring 1,022 images of urban graffiti and tagging captured across Athens. These images of walls, pavements, and statues are utilized to support cultural heritage protection. The data is partitioned into an 80% training and validation set and a 20% testing set, providing a robust framework for developing Deep Learning models.

An automated graffiti detection system is developed for real-time use on mobile devices. Deep Learning and Machine Learning frameworks, such as TensorFlow, are utilized to build the application. Transfer Learning is applied to an Object Detection model to ensure high accuracy and efficiency, allowing graffiti to be instantly identified and processed through a smartphone camera.