A Review of License Plate Recognition Methods Employing Neural Networks

Advances in both parallel processing capabilities because of graphical units (GPUs) and computer vision algorithms have led to the development deep neural networks (DNN) their utilization real-world applications. Starting from LeNet-5 architecture 1990s, modern may tens hundreds layers solve complex problems such as license plate detection or recognition tasks. In this manuscript, we present a review state-of-the-art methods related automatic recognition. Since demonstrated remarkable ability outperform other machine learning techniques, focus only on network based methods. We highlight particular types networks, i.e., convolutional, residual recurrent, long-short-term-memory, used for specific tasks detection, extraction, different existing works. The presented summary also highlights some most widely data sets comparison shares results reported reviewed papers. give an overview effects fog, motion, use synthetic Finally, promising directions future research domain are presented.