The Large Labelled Logo Dataset (L3D): A Multipurpose and Hand-Labelled Continuously Growing Dataset

Asier Gutiérrez-Fandiño1

Barcelona Supercomputing Center

David Pérez-Fernández

Universidad Autónoma de Madrid

Jordi Armengol-Estapé

Barcelona Supercomputing Center

1Corresponding author

Abstract

In this work, we present the Large Labelled Logo Dataset (L3D), a multipurpose, hand-labelled, continuously growing dataset. It is composed of 0.7M 256x256 RGB images extracted from the European Union Intellectual Property Office (EUIPO) open registry. Each of them is associated to multiple labels that classify the figurative and textual elements that appear in the images are associated to them. These annotations have been classified by the EUIPO evaluators using the Vienna classification, a hierarchical classification of figurative marks. We suggest two direct applications of this dataset, namely, logo classification and logo generation.

Paper

Paper

BibTeX

Please cite our work if you use code or data from this site.


@misc{gutierrezfandino2021L3D,
      title={The Large Labelled Logo Dataset (L3D): A Multipurpose and Hand-Labelled Continuously Growing Dataset}, 
      author={Asier Gutiérrez-Fandiño and David Pérez-Fernández and Jordi Armengol-Estapé},
      year={2021},
      eprint={2112.05404},
      archivePrefix={arXiv},
      primaryClass={cs.CV}
}

Data

You can download the data from here or you can download the data using our scripts.

Labelling

How many Vienna codes are assigned by trademark.

How many Trademarks are filled by year.

Sunburst plot containing the distribution of Vienna codes hierarchy by count.