Diseño de un sistema de reconocimiento facial para acceso a áreas restringidas utilizando visión artificial y redes neuronales

Abstract

This project focuses on developing a prototype aimed at enhancing security in restricted access areas. It proposes the creation of a classifier model based on convolutional neural networks, an advanced technique in machine learning that has proven highly effective in classifying complex data. This choice is grounded in the inherent ability of convolutional neural networks to learn patterns and relevant features in datasets, making them an ideal option for addressing security issues such as the identification of authorized and unauthorized individuals in controlled environments. The development process of the prototype involves collecting and preparing input data, as well as implementing suitable training algorithms for the convolutional neural network model. It is expected that this process will enable the model to effectively distinguish between different classes of images, in this case, authorized and unauthorized individuals. Regarding the practical implementation of the system, it involves integrating an artificial vision system that will leverage face detection techniques to capture images of individuals attempting to access the restricted area. These images will be processed by the pretrained model, which will utilize its knowledge to determine whether the subject in question is authorized for access. If affirmative, a signal will be sent to an electric lock via a specific controller, thereby allowing controlled access. A key aspect of implementing this prototype is the use of the Python programming language. Python offers a wide range of libraries and specialized tools in machine learning and artificial vision, significantly facilitating the development and implementation of complex systems like the one proposed in this project. Thanks to Python’s flexibility and robustness, it is expected that the prototype’s development will progress efficiently and effectively, thereby ensuring a robust solution for improving security in restricted areas.