Liashenko A. System of automated syndromic diagnostics by video-laparoscopic images

The object of the study is the process of diagnostics and taking decision by a laparoscopic surgeon when conducting a laparoscopic diagnostic. The aim of the study is to increase the efficiency of diagnostics in laparoscopic examination of abdominal cavity and women`s small bowel organs by developing models, methods and systems for automated syndromatic diagnostic according to the characteristics of video laparoscopic images. Methods of research: methods of discrete information processing, methods of mathematical statistics, determination of informative indicators, methods of computer vision were used for data processing; for the development of decisive rules – the algorithms for the identification of pathological conditions by texture, contour, color characteristics; for training of classifiers – the method of reference vectors and the cascade method; for system and technology development – the methods of system analysis and object-oriented design; for software development – the MS Visual Studio Express environment and C # programming language; to develop an identification and classification system - the OpenCV library; database management system (DBMS) and Microsoft SQL SERVER 2008 R2 environment. Theoretical results: the mathematical model of pathological syndromes detection during laparoscopic diagnostic was developed for the first time. Recognition and identification of pathological objects; classification of their belonging to the corresponding class; processing and diagnostic use the criteria of brightness, color, texture, contour, which allowed to identify on the basis of reliable class structures, some pathological syndromes; calibration method of laparoscopic equipment for the detection of pathological changes in the drug was improved by introducing a preprocessing procedure and taking into account its optical characteristics, which allowed to reduce the effect noise of the image quality, to increase contrast and saturation of the elements, and to ensure high quality of video image segmentation; improved structure of the process of analysis of laparoscopic images by introducing a procedure for monitoring the emergence of the syndromic area on the personnel video sequence in the field of view of the lens of the laparoscopic camera - provides detection and detection of pathological changes in real time; the method of automated diagnostics of diseases by the results of the analysis of laparoscopic video images was further developed by introducing the stage of identification a complex of descriptors characteristic of certain pathological syndromes and development of a diagnostic algorithm, which provides the efficiency of the system of automated syndromic diagnostics by image segmentation as a classifier. Practical results are the development a system of automated syndromatic diagnostics of diseases of the abdominal cavity and small bowel, which allows system to investigate organs and to detect syndromic pathology in the early stages of the disease with an accuracy of up to 90% of all diagnosed cases, and image features by comparing template drugs with the studied and then calculating the difference vector and use of local descriptors. Degree of implementation - the results of the dissertation work were implemented in the practical activity of the Odessa Regional Center of Telemedicine on the base of the Odessa Regional Clinical Hospital and the educational process of the Department of Biomedical Engineering of Vinnitsa National Technical University in the studying of relevant disciplines. Sphere (field) of use - medicine.