| Summary: | <pre>The information technology of recognition and classification of </pre><pre>imaging representations of RCC complicated CKD with use of a </pre><pre>neural network is offered. Approaches to architecture design, </pre><pre>teaching methods, data preparation for training, training and </pre><pre>neural network testing are described. The structural-functional </pre><pre>scheme of the neural network is developed, which consists of </pre><pre>the input, hidden and output layer, each individual neuron is </pre><pre>described by the corresponding activation function with the </pre><pre>selected weights. The expediency of using the number of </pre><pre>neurons, their type and architecture for the task of recognition </pre><pre>and classification of image representations of oncological </pre><pre>phenomena of the organism is shown. Data of patients with </pre><pre>RCC of complicated CKD, research department of reconstructive </pre><pre>and plastic oncourology of NIR, urological department of "Lviv </pre><pre>regional hospital", urology department of Lviv urological regional </pre><pre>medical - diagnostic center, were used as initial data, on the </pre><pre>basis of real observations, a database for training and education </pre><pre>of the neural network was formed. An analysis of the efficiency, </pre><pre>speed and accuracy of the neural network, in particular, a </pre><pre>computer simulation using modern software and mathematical </pre><pre>modeling of computational processes in the middle of the neural </pre><pre>network. Software has been developed for preliminary preparation </pre><pre>and processing of input data, further training and education of the </pre><pre>neural network and directly the process of recognition and </pre><pre>classification. According to the obtained results, the developed </pre><pre>model and structure of the neural network, its software tools </pre><pre>show high efficiency both at the stage of preliminary data processing </pre><pre>and in general at the stage of classification and selection of target </pre><pre>areas of diseases. The next stage of research is the development </pre><pre>and integration of software and hardware system based on </pre><pre>parallel and partially parallel computer technology, which will </pre><pre>significantly speed up computational operations, achieve the learning </pre><pre>and training of the neural network in real time and without loss </pre><pre>of accuracy. The presented scientific and practical results have a </pre><pre>high potential for integration into existing information and analytical </pre><pre>systems, medical analysis the choice of tactics for the treatment </pre><pre>of patients with RCC complicated CKD, and health monitoring </pre><pre>systems in the preoperative and postoperative periods.</pre>
|
|---|
