A NN-based Approach for Monitoring Early Warnings of Risk in Historic Buildings via Image Novelty Detection

Architectural heritage is an important part of the history and identity of countries, but ancient buildings suffer a high vulnerability to hazards, which may induce unpredictable damages. For this purpose, the development of techniques for monitoring historic buildings and immediately alerting in case of early vulnerability assessment is a main objective to be pursued. This paper concerns with a proposal of noninvasive NN-based approach for predicting risk events in artistic buildings. More in detail, a neural approach is suggested for detecting temporal novelties in images of historic evidences with the aim of monitoring early warning of risk events.