A dataset of Romanian (European Union) license plates in VOC format using VOTT.
The dataset is composed of 534 images of which 80% of them are for training and the rest of 20% is for validation. Since the dataset is rather small, it is encouraged to fine-tune a preexisting model with this dataset.
The dataset was shot in both daytime and nightime.