Convolutional Neural Network Method Implementation for License Plate Recognition in Android

TLDR: This study presents the character recognition of vehicle number plates using Convolutional Neural Network (CNN), which is one of the deep learning methods used in the field of number plate recognition using mobile devices.

Abstract: Each vehicle is equipped with an identity in the form of a number plate. Counterfeit documents are often found at the time of examination. Along with the development of artificial intelligence technology, especially in the field of number plate recognition allowing number plate recognition using mobile devices. Using the Android application provides many advantages such as higher recognition accuracy, less resource consumption, and less computational complexity. In this study, the character recognition of vehicle number plates using Convolutional Neural Network (CNN) is one of the deep learning methods. The character recognition process is realized by the segmentation process, which is taking the characters in the number plate. Next is the process of extracting characters with the CNN method. Character extraction results in the form of features. Character features are matched with the pre-prepared character feature database. The test results are very satisfying, which is 94% of the corresponding characters and 6% of characters are not suitable.