The article provides information about the program developed on the basis of the Qt environment, which allows positioning the original image of an object within the field of attention in order to simplify the procedure for generating object features that are invariant to shift, change scale, and rotate its image. Provides an overview of modern methods and software tools for scaling images. The algorithm of the program and a series of computational experiments is described. During the first series, the program positions the image of a triangle within the field of attention using various scaling methods. According to the results of this series, it was concluded which method of scaling an image of an object gives the least loss of quality. In other series of experiments, the program centers and scales the images of a square and a circle inside the attention field with different sizes of the attention field (selection frame) corresponding to a single image scaling factor. Following the results of each series of xperiments, measurements of the sizes of positioned objects were carried out and the dependence of the ratio of their areas on the scaling factor was established. The difference between the maximum and minimum ratio of the coefficients for each series of experiments is calculated. On the basis of the data obtained, it was concluded that for further work with segmented objects of the scene and their positioning in the field of attention, the size of the selection frame of 256x256 pixels can be considered reference.
An improved scheme for image scaling, cropping and colour correction in encrypted domain
