LPRNet

lastmod 2026-11-10created 2019-01-172 edits

LPRNet is a deep learning-based license plate recognition neural network published by Intel researchers. It takes an image of the license plate region as input and reads the character string in one pass using a CNN and CTC loss, without segmenting individual characters, in an end-to-end architecture. It is designed to be lightweight without recurrent neural networks so that it can run in real time on embedded devices. Used together with a license plate detection model, it is applied to the recognition stage of vehicle access control and traffic enforcement systems.

Local Download

LPRNet - License Plate Recognition via Deep Neural Networks

https://arxiv.org/abs/1806.10447

See also

• Deep learning
• License Plate Recognition

Favorite site

Github - Tensorflow based. Use CNN and CTC loss. It is a light network for Plate Recognition