Modelo de clasificación de imágenes violentas basado en un modelo de detección de objetos

Abstract

The present project aims to develop a deeplearning model that allows classifying violent images through the identification of objects, to achieve this a dataset is built with different images in order to determine which objects are considered violent, models are made with different algorithms and/or or deeplearning models which are evaluated in terms of efficiency and it is determined which is the optimal one to use. This project is part of the base of the master's thesis that consists of a software model for managing cultural images based on Multiview Learning where the model developed in this work is used as a module of the Multiview Learning model, in addition the master's thesis composes part of the doctoral work on a navigation model using computational intelligence techniques to improve the experience in the visualization of historical and cultural data.