ESTER-Pt: An Evaluation Suite for TExt Recognition in Portuguese

Optical Character Recognition (OCR) is a technology that<br> enables machines to read and interpret printed or handwritten texts from<br> scanned images or photographs. However, the accuracy of OCR systems<br> can vary depending on several factors, such as the quality of the input<br> image, the font used, and the language of the document. As a general tendency, OCR algorithms perform better in resource-rich languages as they<br> have more annotated data to train the recognition process. We propose ESTER-Pt, an Evaluation Suite for TExt Recognition in Portuguese in this work. Despite being one of the largest languages in terms of speakers, OCR in Portuguese remains largely unexplored. Our evaluation suite<br> comprises four types of resources: synthetic text-based documents, syn-<br> thetic image-based documents, real scanned documents, and a hybrid set<br> with real image-based documents that were synthetically degraded.