Research of methods of distribution of graphic objects on websites for assessment of interface quality

Abstract

Formulation of the problem. An actual problem when creating web resources is the test-ing of the designed design, which is the direction of testing the user interface. This paper con-siders ways to automate the analysis of the interface of web pages in terms of visual perception of man. The aim of the work. The aim of the work is to improve the quality of automated testing of the graphical user interface based on the use of methods of analysis of color digital images and detection of graphic objects. Methods of selecting objects on digital images. The traditional method of segmentation is described – threshold binarization, which results in a binary image. The JSEG algorithm is more advanced. According to this method, instead of estimating the parameters of the texture model, the homogeneity of each image fragment is checked, which leads to a reduction in the amount of computation. Website interface evaluation methods. In order to qualitatively evaluate the user interface, it is necessary to identify the basic principles on which designers rely when designing interfaces. Basic principles of interface construction: shape, size, brightness, color, direction, location. Determining the brightness characteristics of the image. To estimate the brightness pa-rameters, you can use the estimation of the image histogram, which displays the brightness value. The optimal type of brightness histogram is the normal distribution. Determining the number of primary colors in the image. Performed by constructing a histogram for an indexed image on a given color map. It is optimal to use no more than 3 primary colors. Selectionobjects in the image. The methods of image segmentation described above are implemented and the imperfection of these methods is shown. The application of the JSEG algorithm gave a more positive result, but requires additional adaptation to our tasks. Conclusions. The study and comparative analysis of digital image processing methods to automate the process of assessing the quality of the graphical user interface. It is concluded that the application of the classical approach to image segmentation did not give the desired result for image analysis of web pages, namely there are problems of inability to separate adjacent graphics at the threshold binarization, and combine text characters into one block. The JSEG algorithm includes color quantization and spatial segmentation operations, due to which the result of object selection is better, but still requires further settings.