✓ In this web page, a novel comprehensive database for conducting research on offline Persian handwritten recognition is introduced. 500 native Persian writers who were equally balanced in terms of gender (250 men+250 women) and randomly selected, filled 7 different pages of handwritten sample forms (HSFs). The filled forms (500*7=3500 HSFs) were scanned at 300 DPI resolution, and then all their handwritten items were extracted, segmented, organized, grouped, and labelled as an structured database. This database includes images of a huge number of handwritten isolated digits, numeral strings, touching digits, dates, words, names, alphabetical letters, free texts, arithmetic and especial symbols. For each sample of this database, writer ID, gender, age, handedness, and education level of its writer has been provided, and all samples in this database were assigned with detailed ground truth and stored in 3 color formats: true color, gray level, and binary; also all subsets of this database were randomly partitioned into training, validation and testing sets to be ready for pattern recognition, machine learning, and image processing experiments, researches, and investigations.
This database is freely available to the research community and for research purposes. A full professional version is freely available upon receiving official request and registrations for all researchers in this field. Please send your requests or comments to:
Dr. Javad Sadri at: j_sadri [at] encs.concordia.ca or to: Mr. Mohammad Reza Yeganehzad at: m.yeganehzad [at] gmail.com
This database has been completely explained in the following research paper:
If you are using this database, please cite the above paper. We hope this comprehensive database will extend research in pattern recognition community.
✓ A snapshot (very small version in gray level format) of this database is publicly available and can be downloaded here: "Please click here to download". A sophisticated web site soon will be created for this database for accepting requests and downloading this database in full and also to report other developments, updates, news, comparisons, contests and competitions, and also published papers based on this database.
✓ A summary of statistics of items in this database is shown below:
✓ Few black and white samples of worded dates, numeral strings, words, and texts taken from our database are shown below:
Javad Sadri, Mohammad Reza Yeganehzad, Javad Saghi, "A Novel Comprehensive Database for Offline Persian Handwriting Recognition", Published in the Journal of "Pattern Recognition (Elsevier)", Volume 60, Pages 378-393, December 2016. Currently available on-line at: http://www.sciencedirect.com/science/article/pii/S0031320316300097.