Deep Learning-Based Text Recognition of Agricultural Regulatory Document

In this study, an OCR system based on deep learning techniques was deployed to digitize scanned agricultural regulatory documents comprising of certificates and labels. Recognition of the certificates and labels is challenging as they are scanned images of the hard copy form and the layout and size of the text as well as the languages vary between the various countries (due to diverse regulatory requirements). We evaluated and compared between various state-of-the-art deep learning-based text detection and recognition model as well as a packaged OCR library - Tesseract. We then adopted a two-stage approach comprising of text detection using Character Region Awareness For Text (CRAFT) followed by recognition using OCR branch of a multi-lingual text recognition algorithm E2E-MLT. A sliding windows text matcher is used to enhance the extraction of the required information such as trade names, active ingredients and crops. Initial evaluation revealed that the system performs well with a high accuracy of 91.9% for the recognition of trade names in certificates and labels and the system is currently deployed for use in Philippines, one of our collaborator's sites.