MLT Image Text Recognition Dataset

The MLT dataset is provided by the Robust Reading Challenge and contains 10,000 image files and the location information and related text annotations of the text in each image. The dataset includes 10 major languages and can be used for research on text recognition and extraction in images.