A high-quality, balanced dataset of 330,000 images featuring various types of Chinese license plates. The dataset is generated using Generative Adversarial Networks (GANs), ensuring excellent image quality and a balanced distribution of different license plate types. This dataset is perfect for training and evaluating license plate recognition models.
Note: In order to ensure the balance of the characters, there may be some license plate numbers that do not conform to common sense (for example, all letters), but this phenomenon does not affect the training.
Paper | Code | Results | Date | Stars |
---|