IDPL PFOD Dataset

IDPL-PFOD, and its larger successor IDPL-PFOD2, are large-scale datasets designed to advance Optical Character Recognition (OCR) for the Farsi language, addressing the challenges posed by its cursive script and diacritics. Research focuses on evaluating the effectiveness of various deep learning architectures, including Convolutional Recurrent Neural Networks (CRNNs) and Vision Transformers, for accurate Farsi text recognition using these datasets. The development of such datasets is crucial for improving OCR technology in under-resourced languages and has broader implications for document processing and information retrieval. Similar large-scale datasets are being developed for other applications, such as crop-weed recognition in precision agriculture and analyzing criminal justice data for fairness studies, highlighting a growing need for comprehensive, high-quality datasets to drive advancements in machine learning across diverse fields.