Аннотации:
© Springer Nature Singapore Pte Ltd. 2019. Intelligent search in the digital libraries is very important. It is very important for efficient research to obtain relevant information quickly. In the paper, we propose methods for automatic processing of online resources for the institution library using scanned copies or/and pdf files to make MathML model and provide extended search capacity. The key idea is to use Thinking–Understanding framework to provide automatic document-type detection and processing using the thinking flow to combine different open-source engines like OCR and approaches like Word2vec.