Semi-automated OCR database generation for Nabataean scripts

A large amount of real-world data is required to train and benchmark any character recognition algorithm. Developing a page-level ground-truth database for this purpose is overwhelmingly laborious, as it involves a lot of manual efforts to produce a reasonable database that covers all possible words of a language. Moreover, generating such a database for historical (degraded) documents or for a cursive script like Urdu¹ is even more complex and grueling. The presented work attempts to solve this problem by proposing a semi-automated technique for generating ground-truth database. It is believed that the proposed automation will greatly reduce the manual efforts for developing any OCR database. The basic idea is to apply ligature-clustering prior to manual labeling. Two prototype datasets for Urdu script have been developed using the proposed technique and the results are also presented.