Information about the data:
Standard datasets play an important role in development, evaluation and comparison of different recognition systems. In this paper, we present a new dataset called SRU/PHN for Farsi offline handwritten recognition. The SRU/PHN dataset contains 22000 handwritten Farsi names with samples extracted from 280 forms which are filled by 280 writers with different ages, genders and education levels. For each image in the dataset, the ground truth information including corresponding image file, image size, number of sub-words, characters dots are also available. The SRU/PHN dataset is designed for training and testing recognition systems for handwritten Farsi words and is publically available for research purposes.
please enter your information below; then, the database will be available to download.