Field layout analysis for Russian forms

摘要

This article is about the structural analysis of the text in OCR and form processing systems. The authors propose universal approach for layout understanding in the text lines, equally suitable for unstructured typewritten sheets and fields of documents (forms). Developed algorithm is robust to distortions, typical for flatbed scanners as well as for cameras, incorporated to mobile devices. It was implemented in C++ and integrated to Cognitive Forms 2.0 system. The methodology of this work is based on auto-classification, cluster analysis and histogram analysis. This article is focused on precise estimating of text line positions and word finding in forms.