@InProceedings{LarocaSanEstLuzMen:2022:FiLoDa,
               author = "Laroca, Rayson and Santos, Marcelo and Estevam, Valter and Luz, 
                         Eduardo and Menotti, David",
          affiliation = "{Federal University of Paran{\'a}} and {Federal University of 
                         Paran{\'a}} and {Federal University of Paran{\'a}} and {Federal 
                         University of Ouro Preto} and {Federal University of 
                         Paran{\'a}}",
                title = "A First Look at Dataset Bias in License Plate Recognition",
            booktitle = "Proceedings...",
                 year = "2022",
         organization = "Conference on Graphics, Patterns and Images, 35. (SIBGRAPI)",
             keywords = "license plate recognition, dataset bias.",
             abstract = "Public datasets have played a key role in advancing the state of 
                         the art in License Plate Recognition (LPR). Although dataset bias 
                         has been recognized as a severe problem in the computer vision 
                         community, it has been largely overlooked in the LPR literature. 
                         LPR models are usually trained and evaluated separately on each 
                         dataset. In this scenario, they have often proven robust in the 
                         dataset they were trained in but showed limited performance in 
                         unseen ones. Therefore, this work investigates the dataset bias 
                         problem in the LPR context. We performed experiments on eight 
                         datasets, four collected in Brazil and four in mainland China, and 
                         observed that each dataset has a unique, identifiable 
                         {"}signature{"} since a lightweight classification model predicts 
                         the source dataset of a license plate (LP) image with more than 
                         95% accuracy. In our discussion, we draw attention to the fact 
                         that most LPR models are probably exploiting such signatures to 
                         improve the results achieved in each dataset at the cost of losing 
                         generalization capability. These results emphasize the importance 
                         of evaluating LPR models in cross-dataset setups, as they provide 
                         a better indication of generalization (hence real-world 
                         performance) than within-dataset ones.",
  conference-location = "Natal, RN",
      conference-year = "24-27 Oct. 2022",
                  doi = "10.1109/SIBGRAPI55357.2022.9991768",
                  url = "http://dx.doi.org/10.1109/SIBGRAPI55357.2022.9991768",
             language = "en",
                  ibi = "8JMKD3MGPEW34M/47M827H",
                  url = "http://urlib.net/ibi/8JMKD3MGPEW34M/47M827H",
           targetfile = "laroca2022first-inpe.pdf",
        urlaccessdate = "2024, Apr. 29"
}