This paper is an extended description of one of the basic models of descriptive image analysis that characterizes the architecture and structure of the image recognition process: a multi–level model of image analysis and recognition procedures based on the joint use of methods for combining algorithms and methods for combining fragmentary image data – partial descriptions of the object of analysis and recognition - images. The architecture, functionality, limitations and characteristics of a multilevel model for combining algorithms and initial data in image recognition are substantiated and defined. Scenarios of application of a multilevel model of image analysis and recognition procedures are presented using the example of solving the problem of analyzing images obtained using optical coherence tomography angiography to automate the detection of pathological changes in the morphometric characteristics of the fundus.