1 dataset found

Groups: Text Localisation

Filter Results

  • SynthText

    SynthText dataset is proposed by Gupta et al. for scene text detection. The original dataset is composed of 800,000 scene text images, each with multiple word instances.