2026-04-17T17:17:49Zhttps://recercat.cat/oai/request

oai:recercat.cat:2117/3637832025-07-22T15:52:40Zcom_2072_1033col_2072_452951

Anti-spoofing mechanisms in face recognition Research, implementation and comparison between methods for pupil detection in an image Bové Escribano, Sergi Universitat Politècnica de Catalunya Slovenská technická univerzita v Bratislave Sopiak, Dominik Àrees temàtiques de la UPC::Informàtica::Intel·ligència artificial Artificial intelligence Deep learning Neural networks (Computer science) Human face recognition (Computer science) suplantació d'identitat detecció de pupila reconeixement facial intel·ligència artificial aprenentatge automàtic aprenentatge profund xarxes neuronals anti-spoofing pupil detection face recognition artificial intelligence machine learning deep learning neural networks Intel·ligència artificial Aprenentatge profund Xarxes neuronals (Informàtica) Reconeixement facial (Informàtica) The objective of this work is to design and implement a pupil detection solution to be used as part of an anti-spoofing mechanism in face recognition. This is intended to improve the security and reliability of face recognition technology by minimizing its risks and, thus, provide confidence to users and companies that employ it. In addition, the project wants to focus on the efficiency and accuracy of the final solution, to do so, various alternatives based on different methods will be developed and compared. To achieve this, the project follows a linear work methodology, that is to say, starting from the search of information, through the theoretical design phase, the code implementation and ending with the presentation of an operational solution. The results achieved after the whole process have been satisfactory and show the viability of the target set. Also, the work provides a series of possible future improvements while reflecting how the initial idea has been evolving and maturing to end up resulting in a product more solid and adjusted to the detected need. 2020-09-02 Master thesis https://hdl.handle.net/2117/363783 152738 eng Restricted access - author's decision application/pdf Universitat Politècnica de Catalunya