Bezpečnostní řešení kamerové detekce za využití Raspberry Pi

Abstract

This masters thesis deals with the construction of a camera system based on a single-board computer from the Raspberry Pi Foundation with intelligent face recognition. The thesis deals with the theory of biometrics together with machine learning. Furthermore, the author presents the means used in his work, from hardware components to used software elements such as detectors or classifiers. The next part of the work is devoted to experiments where the author tested reliability of selected algorithms. At the end of the work is a financial comparison of camera system against a commercial product. The created camera system is less demanding than a commercial competitor, however, the results of experiments show a weaker face detection of the algorithm in night mode.

Description

Subject(s)

Raspberry Pi, machine learning, deep learning, face recognition, single-board computer, neural network, VSS, security camera

Citation