Development of the architecture of an information and analytical image processing system

Abstract

The paper proposes the architecture of an information and analytical system for organizing reproducible research in the field of computer vision. The system provides unified collection and storage of visual data from heterogeneous sources, with the ability to automatically perform analytical tasks. Their processing is carried out through integration with arbitrary algorithms - both classical methods of image analysis and neural network models within a single execution interface. The platform implements a mechanism for dynamic management of the context of experiments and provides built-in tools for comparative analysis of the results of various algorithms. The developed solution formalizes the full research cycle, reducing the routine load and ensuring full reproducibility when working with visual data.