Robust license plate recognition using neural networks trained on synthetic images

作者：

Highlights：

• A complete vehicle license plate reading system is proposed.

• CNNs trained for LPR on only synthetic samples generalizes well to real images.

• A re-classification of fully convolutional classifications improves performance.

• Results on other datasets shows an improvement over previous LPR systems.

摘要

•A complete vehicle license plate reading system is proposed.•CNNs trained for LPR on only synthetic samples generalizes well to real images.•A re-classification of fully convolutional classifications improves performance.•Results on other datasets shows an improvement over previous LPR systems.

论文关键词：License plate recognition (LPR),Convolutional neural network (CNN),Synthetic training

论文评审过程：Received 23 August 2018, Revised 19 March 2019, Accepted 9 April 2019, Available online 10 April 2019, Version of Record 22 April 2019.

论文官网地址：https://doi.org/10.1016/j.patcog.2019.04.007