License plate image character generation by conditional variational auto-encoders for license plate recognition | |
| Author | Pattranit Teerakoson |
| Call Number | AIT Thesis no.ISE-24-09 |
| Subject(s) | Automobile license plates--Data processing Image processing Generative adversarial networks (Computer networks) Deep learning (Machine learning) |
| Note | A thesis submitted in partial fulfillment of the requirements for the degree of Master of Engineering in Mechatronics and Machine Intelligence |
| Publisher | Asian Institute of Technology |
| Abstract | License Plate Character Recognition (LPCR) is a system that reads automobile registration plates using optical character recognition from photos and videos, and it has a lengthy history due to its use. While deep learning has made great advances in LPCR, training deep networks for the LPCR module necessitates a huge number of license plate (LP) images and comments. Unlike other public datasets of vehicle information, each LP has a distinct collection of characters and digits based on the country or location. As a result, gathering enough LP photos is quite challenging for standard study. In this research, we offer Conditional Various Autoencoder (C-VAE), a Thai-LP characters image creation method, using an ensemble of Various Autoencoder (VAE), Additionally, we offer an effective end-to-end LPCR module using a modified lightweight YOLOv8 model. Hundreds of synthetic LP images were produced using C-VAE, as there are only 200 real LP photos available online. The generated pictures were found to be useful for training the LPCR module in addition to looking like actual ones. Lastly, character alterations on license plates might add to character databases that are lacking. When compared to 500 produced images and 200 real photos, it obtains an accuracy grade of 0.7539.They found that utilizing augmentation to create generated pictures might be like adding a dataset. |
| Year | 2024 |
| Type | Thesis |
| School | School of Engineering and Technology |
| Department | Department of Industrial Systems Engineering (DISE) |
| Academic Program/FoS | Mechatronics and Machine Intelligence (MMI) |
| Chairperson(s) | Mongkol Ekpanyapong; |
| Examination Committee(s) | Chaklam Slipasuwanchai;Huynh, Trung Luong; |
| Scholarship Donor(s) | His Majesty the King’s Scholarships (Thailand); |
| Degree | Thesis (M. Eng.) - Asian Institute of Technology, 2024 |