Recognition Module for the Identification and Classification of Patterns in Vehicles Using Deep Learning Techniques

Abstract

In Mexico, vehicle theft is a serious problem that affects a considerable number of citizens. In the state of Veracruz, 61,796 cars have been reported stolen in one year. This paper presents a pattern recognition system using Artificial Intelligence techniques. This system uses video image analysis and a Deep Learning module to classify vehicles into two categories: automobiles and motorcycles. A case study used video surveillance cameras to capture more than 1,000 vehicles. The Deep Learning model achieved 85% classification accuracy. This solution will improve the identification of suspicious and stolen cars, contributing to security and traffic control.