Abstract. 

In this paper, the input data used in modern identification document recognition systems are considered in detail. The classification of recognition systems according to the type of input data being processed is proposed. The existing problems and features of defined 2D, 3D and 4D recognition are considered in detail. Conceptual model and approaches for the construction of a universal industrial recognition system capable of processing input data belonging to different classes in the same way are proposed.

Keywords: 

document recognition, identification documents, artificial intelligence, OCR, machine learning, image processing

PP. 3-9.

DOI: 10.14357/20790279220301