Claudia-Ioana Satnoianu defended his Master Thesis on Face Recognition, Effects and Detection of Beautification Applications

Claudia-Ioana Satnoianu defended her Master Thesis on Face Recognition, Effects and Detection of Beautification Applications

The goal of this MSc thesis is to analyze the impact beautification apps can have on face recognition systems and to see if detection systems are capable of noticing when an image has been beautified. The MSc thesis is composed of seven chapters, each of them dealing with a different aspect of the above-mentioned topic. An introduction is made in Chapter 1, where the reader gets familiarized with topics such as facial image databases, face recognition systems and detection systems. Chapter 2 is divided into two parts, the first one of them presenting a literature review and the second one proposing the goals which are set for this thesis project. Chapter 3 focuses on presenting the entire process of selecting the beautification apps which are being used during the project based on a set of criteria, to how to use the apps and then what are the final results that can be achieved by using them. Chapter 4 concentrates on both the bona fide databases that are used and on how the beautification process of this bona fide images can be automated, for a faster database generation procedure. After generating the database that serves the scope of this project, in Chapter 5a description of the detection systems is made. The chapter is split in two, the first part focusing on the possible scenarios a detection system can be useful, whereas the second part focuses on the detection systems that are being used when doing the experiments. Chapter 6 is divided into three parts. Part one describes how the experiments for both face recognition systems and detection systems were conducted, whereas parts two and three focused on analyzing the experiments‘ results for each type of system. Conclusions are drawn in Chapter 7. The main aim of the MSc thesis has been reached. The author suggests that face recognition systems should be improved so that they have a higher performance rate given the beautification scenarios.