Huralnyk A. Method and ultrasound tool for assessment of the condition of the hip joint

The object of investigation is the processes of processing ultrasound images of the hip joint. The aim of the investigation is to increase the reliability of the assessment of the condition of the hip joint by developing new models and methods of analysis of ultrasound images of the hip joint and automatic hardware and software based on them. Research methods: mathematical apparatus of discrete signal theory, digital image representation for development of theoretical principles of speckle noise filtering and methods of spatial image processing for development of speckle noise attenuation filter; measurement theory to assess the metrological characteristics of the proposed methods and systems; methods of mathematical modeling using modern computer technology. Theoretical results: for the first time a method of automatic assessment of hip condition based on 2D ultrasound images with classification of hip condition by rejection of images of incorrect hip position, visualisation of its anatomical landmarks and measurement of angular parameters was proposed, which increased the reliability by 10-15 %; The method of filtering ultrasonic images has been improved, which differs from the known sequence of morphological transformations, which allowed to take into account the level of noise and intensity differences, and accordingly change the filtration depth and obtain higher quantitative indicators of filtration quality 7% higher than known); further developed a mathematical model for the selection of a given fragment on the ultrasound image by its spectral features, which in contrast to the known uses the equivalence function, which allowed to obtain more accurate coordinates of reference points (relative error is 6.25%), which is the formation of lines, and hence the angles for classifying the condition of the hip joint; For the first time, an automatic method for determining a specific 2D ultrasound image frame corresponding to the correct position of the ultrasound head on the child's body was proposed by applying adaptive neural network analysis, which allowed to reject false images and minimise joint position uncertainty with further processing of the correct imaging frame. Practical results are the development of hardware and software based on the evaluation of geometric parameters obtained from ultrasound images of the hip joint, as well as software for automatic rejection of images, filtering and visualisation of anatomical landmarks with measurement of angular parameters of the hip joint, which can be used -diagnostic centers in the study of diseases of the hip joint of children under three months. Degree of implementation - the results of the dissertation are implemented on the basis of the Clinic of Pediatric Surgery of Vinnytsia National Medical University. E. Пирогова; in the educational process of Vinnytsia National Technical University in teaching the relevant disciplines. Sphere (branch) of use - medicine.