Smart Home: Deep Learning as a Method for Machine Learning in Recognition of Face, Silhouette and Human Activity in the Service of a Safe Home

George Vardakis, George Tsamis, Eleftheria Koutsaki, Kondylakis Haridimos, Nikos Papadakis

Despite the general improvement of living conditions and the ways of building buildings, the sense of security in or around them is often not satisfactory for their users, resulting in the search and implementation of increasingly effective protection measures. The insecurity that modern people face every day, especially in urban centers regarding their home security, led computer science to the development of intelligent systems, aiming to mitigate the risks and ultimately lead to the consolidation of the feeling of security. In order to establish security, smart applications were created that turned a house into a Smart and Safe Home. We first present and analyze the deep learning method and emphasize its important contribution to the development of the process for machine learning, both in terms of the development of methods for safety at home, but also in terms of its contribution to other sciences and especially medicine where the results are spectacular. We then analyze in detail the back propagation algorithm in neural networks in both linear and non-linear networks as well as the X-OR problem simulation. Machine learning has a direct and effective application with impressive results in the recognition of human activity and especially in face recognition, which is the most basic condition for choosing the most appropriate method in order to design a smart home. Due to the large amount of data and the large computing capabilities that a system must have in order to meet the needs of a safe, smart home, technologies such as fog and cloud computing are used for both face recognition and recognition of human silhouettes and figures. These smart applications compose the systems that are created mainly through “Deep Learning” methods based on machine learning techniques. Based on the study we have done and present in this work, we believe that with the use of DL technology, the creation of a completely safe house has been achieved to a large extent today, covering an urgent need these days due to the increase in crime.