The aim of this study was to analyze the possibility of using a mathematical model of logistic regression for the recognition of malignant neoplasms on digital images of the skin. The study used a database containing 6594 digital skin images. At the first stage of the study, digital skin images were segmented to isolate the object under study, in which morphometric and color characteristics corresponding to the parameters of the classical ABCD system were determined. At the second stage, the characteristics were used in the classification into malignant and benign neoplasms using logistic regression. When classifying images, the highest value of the accuracy indicator (67.9 [66.9; 68.8]%) was obtained during classification using logistic regression, built on the basis of the reverse Wald stepwise method. Thus, the logistic regression built on the basis of the reverse stepwise Wald method can be applied in the classification of malignant neoplasms on digital images of the skin, but further research and determination of the optimal parameters are required.

Keywords: mathematical model, digital skin images, logistic regression, image classification, skin cancer

The purpose of this research is analysis of the possibility of using classical neural networks for the malignant neoplasms recognition on skin digital images. In this study was used International Skin Imaging Collaboration database including 6594 dermatoscopic images. At the first stage of the study, digital skin images were classified into malignant and benign neoplasms using the IBM SPSS Statistic tool to automatically select the architecture for artificial neural network. At the second stage was built architecture of artificial neural network, which include one hidden layer. At the third stage was used architecture of an artificial neural network with two hidden layers. In the course of the study digital skin images classified with the highest value of the accuracy indicator (0,752 [0,736 ; 0,768]) during classification using the architecture of an artificial neural network, which includes two hidden layers. The sigmoid was used as the activation function for the hidden layers. The hyperbolic tangent was used as activation function for output layer. With this value of the accuracy, the specificity of the diagnostic method was obtained – 0,813 [0,802 ; 0,824], as well as the value of the sensitivity – 0,665 [0,637 ; 0,691]. Thus, artificial neural networks can be used as a method for skin malignant neoplasms diagnostic on digital images.

Keywords: artificial neural networks, digital skin images, machine learning, image classification, skin cancer

The article considers the application of criteria for assessing the quality of segmentation of digital images of sputum stained by the method of Ziehl–Neelsen to select the optimal parameter "Sigma" wavelet transform Mexican Hat. 830 digital images obtained by sputum smear microscopy were used as the study material. To assess the optimal selection of the parameter σ, we used the average number of objects selected in the images, the proportion of missed acid-resistant mycobacteria in the images, the homogeneity criterion and 3 complex criteria for assessing the quality of image segmentation. The analysis showed that with an increase in the parameter σ there is a slight decrease in the value of the homogeneity criterion. At the same time, the parameter σ increases from 2.4 and more according to the complex criteria, and the image segmentation quality improves. Thus, the most optimal values of the σ parameter of the Mexican Hat wavelet for segmentation of digital images of sputum stained by the cyl-Nielsen method are values in the range from 2.90 to 3.09.

Keywords: method Ziehl-Nielsen, image segmentation, quality evaluation criteria, wavelet transform, Mexican Hat

The comparison of different methods of selection of signs for identification of objects on digital images of microscopic preparations of sputum, colored by the method of tsilya-Nilsen. The following methods were considered: the method of intersections, Shannon, kulbaka and accumulated frequencies. It is concluded that the method of intersections allows the selection of features from the entire feature space so that the classification models allow to obtain the maximum accuracy of classification with the least number of input parameters.

Keywords: method Ziehl-Nielsen, object recognition, image recognition, selection of features, the method of cumulative frequencies, the method of Shannon, Kullback, the method of intersections, logistic regression, classification tree, discriminant analysis

A comparison of different methods of segmentation of digital images of sputum stained by the method of Ziehl-Nielsen. We considered the following methods: threshold binarization, method binarization Otsu, detectors borders (operators Roberts, Sobel, Prewitt, Robinson and Kenny), detectors of Harris corners and FAST (Features from Accelerated Segment Test) algorithm, artificial neural network and wavelet transform Mexican Hat, as well as the search function of the contours of the OpenCV library. To analyze the quality of the image segmentation and time spent for carrying out segmentation. Concluded that the use of the wavelet transform Mexican Hat has the best quality segmentation with a relatively small time spent.

Keywords: the method of Ziehl-Nielsen, segmentation, digital imaging, detector angles, FAST, operator Kenny, the Sobel operator, Roberts operator, the operator Prewitt, operator Robinson, artificial neural networks, OpenCV