COCO-Text API http://vision.cornell.edu/se3/coco-text/
COCO-Text is a large dataset designed for text detection and recognition. This is a Python API that assists in loading, parsing and visualizing the annotations. The format of the COCO-Text annotations is also described on the project website.
In addition to this API, please download both the MSCOCO images, available on the [MSCOCO project website] (http://mscoco.org/dataset/#download) and the text annotations from the [coco-text website] (http://vision.cornell.edu/se3/coco-text/).
This dataset is based on Microsoft COCO. Please visit http://mscoco.org/ for more information on COCO, including the image data, object annotatins and caption annotations.
After downloading the images and annotations run the Python demo for example usage.