CNN-RNN based method for license plate recognition

• Published in: CAAI Transactions on Intelligence Technology 2018-09, Vol.3 (3), p.169-175
• Main Authors: Shivakumara, Palaiahnakote, Tang, Dongqi, Asadzadehkaljahi, Maryam, Lu, Tong, Pal, Umapada, Hossein Anisi, Mohammad
• Format: Article
• Language: English
• Subjects: (B6135E) Image recognition; (C5260B) Computer vision and image processing techniques; (C5290) Neural computing techniques; Artificial neural networks; bi‐directional long‐short term memory; BLSTM; Classification; classification methods; CNN‐RNN based method; context information; convolutional neural networks; cursive handwriting; dark background; Deep learning; dense cluster‐based voting; feature extraction; Feature recognition; foreground colour; high discriminative ability; image classification; image colour analysis; image recognition; image segmentation; learning (artificial intelligence); license plate images; license plate recognition; License plates; Malaysia; Malaysian government; Methods; MIMOS; multiple adverse factors; Neural networks; Noise; public vehicles; recognition performance; recurrent neural nets; Recurrent neural networks; Special Issue: Selected Papers from The 4th Asian Conference on Pattern Recognition (ACPR 2017); tandard dataset UCSD; Text categorization; Vehicle identification; vehicle movements; white background
• ISSN: 2468-2322; 2468-2322
• DOI: 10.1049/trit.2018.1015

Achieving good recognition results for License plates is challenging due to multiple adverse factors. For instance, in Malaysia, where private vehicle (e.g., cars) have numbers with dark background, while public vehicle (taxis/cabs) have numbers with white background. To reduce the complexity of the problem, we propose to classify the above two types of images such that one can choose an appropriate method to achieve better results. Therefore, in this work, we explore the combination of Convolutional Neural Networks (CNN) and Recurrent Neural Networks namely, BLSTM (Bi-Directional Long Short Term Memory), for recognition. The CNN has been used for feature extraction as it has high discriminative ability, at the same time, BLSTM has the ability to extract context information based on the past information. For classification, we propose Dense Cluster based Voting (DCV), which separates foreground and background for successful classification of private and public. Experimental results on live data given by MIMOS, which is funded by Malaysian Government and the standard dataset UCSD show that the proposed classification outperforms the existing methods. In addition, the recognition results show that the recognition performance improves significantly after classification compared to before classification.