A methodology for assessing the quality of software products that include artificial intelligence in the tasks of document classification and recognition

Abstract

The article presents the results of research conducted by the Russian IT company “Logstrim” to assess the quality of software products that include artificial intelligence in document classification and recognition tasks. During the research, a methodology was developed that allows us to evaluate the quality of software products that include artificial intelligence in document classification and recognition tasks. The developed methodology was used to evaluate the quality of the “ECO-DOC” software package, which includes the YOLO v5 neural network service, in document classification and recognition tasks. The results of testing and quality assessment of the ECO-DOC software package are presented.

Keywords

• methodology
• software package quality assessment
• neural network
• artificial intelligence system
• “eco-doc” software package
• quality characteristics
• quality indicators
• quality indicator elements
• functional correctness metrics