On Modeling Descriptive Image Analysis Procedures on a Specialized Turing Machine
Igor B. Gurevich, Vera V. Yashina · Pattern Recognition and Image Analysis · 2022
Abstract This article is devoted to an original approach to the definition and description of a Turing machine (TM) for implementing descriptive image analysis methods based on an information structure for generating descriptive algorithmic schemes for automating image analysis. The fundamental problem, to which the subject of the study belongs, is the automation of extracting information from images that is necessary for making intelligent decisions. One of the important and promising areas of research in this problem is the automation of the choice of a method for solving the problem of image analysis. A necessary condition for such automation is a comparative analysis and optimization of the image analysis algorithms, which, in turn, requires estimates of the complexity and efficiency of algorithms and a universal calculator to obtain them. One of the strategic goals for the development of descriptive image analysis is the study of models of image analysis processes. To do this, it is proposed to define and build an Image Analysis Machine, i.e., TMs specialized for processing spatial information. A method for determining a TM for modeling descriptive algorithmic schemes for image analysis is proposed and described. This machine can also be used to evaluate the mathematical characteristics of image analysis algorithms. The main concepts and objects of descriptive analysis necessary for defining and constructing a TM for modeling and studying procedures for the descriptive analysis of images are recalled. An example is given of modeling, on a TM specialized for image analysis, information processing procedures implemented in an information structure to generate descriptive algorithmic schemes for image analysis when solving image analysis problems. The fundamental importance of the results of these studies for the development of the mathematical theory of image analysis and their scientific novelty are related to the formulation of problems and the development of methods for modeling the processes of automating image analysis when using poorly formalized image representations as the initial data, including their own spatial data: images and their fragments, image models, incompletely formalized representations, and subsets of combinations of these representations. The idea of using a TM to model and explore the complexity of descriptive analysis algorithmic procedures is completely original. This example testifies to the productivity of the method proposed by the authors for determining the TM for image analysis.