Abstract
Image recognition occupies a special place in the science of computer vision. Contour analysis particular important in pattern recognition. In this paper, we consider the structure of the contour analysis of external images. Active contour method shows the possibilities of using the minimum energy curve. The energy function that defines boundaries without first isolating the boundaries of an image object. Therefore, the Canny Boundary Detector algorithm is used to detect the contours of an object in an image. This algorithm smooths out image blur and eliminates noise, eliminates errors or interference in the picture. The path tracking method crosses out the boundaries between the subject and the background. Algorithms such as machine learning are needed in order to reflect its performance and the need for recognition. Clustering is used as a machine learning method to find the nearest neighbor criteria. The mathematical model of clustering has its own uniqueness and application for identifying borders or contours in the image. The contour detection and linking approach uses graph analysis, which works and does not lose efficiency in the presence of noise. The convexity defects contour analysis algorithm is able to determine the size of borders, and also uses the search for contour recesses and the introduction of key features when calculating object parameters.