Pisemne opisy obrazów wykonywanych przez diagnostę podczas badań endoskopowych niekiedy nie wystarczają do prawidłowego prowadzenia przez lekarza procesu leczenia. Obecnie oprócz opisu do dokumentacji medycznej dodaje się wybrane zdjęcia wykonane podczas badań endoskopowych. Ważne jest, aby zdjęcia te charakteryzowała ich wysoka czytelność. W artykule została przedstawiona metoda wstępnej obróbki materiału wideo umożliwiająca przeprowadzenie procesu klasyfikacji i segregacji.
EN
Sometimes the text descriptions of the endoscopic images for a doctor are insufficient to correct lead the treatment process. At preset time, moreover text descriptions the endoscopic images enter into the composition of the medical documentations. Is very important in order to placed images were characterized high readable. The paper presents the method of the introductory images processing make a classification and segregation process possible.
The currently observed dramatic development of computer methods supporting diagnostic processes in various branches of medicine, and the development of methods of digital processing of images insinuates the creation of ever more sophisticated programmes supporting the work of medical doctors. The following paper deals with the problem of automatic determination of the edges of defect lesions in the alimentary tract, and presents suggestions as to creation of certain numerical parameters that would objectivise the nature of the edge of the lesion examined, thus contributing new information aimed to help the doctor make a diagnosis based on endoscopic image of the lesion. An important issue in that respect is that the image of the alimentary tract mucous membrane is highly irregular, the fact stemming from general anatomic and histological structure of mucosa of the alimentary tract. The pictures, prior to further processing and analysis, require the application of a smoothing filter, sometimes even a few times. Unfortunately, simple smoothing (averaging) filters cause considerable distortion of edges of the objects, like blurring, flattening and smoothing of the edges. With computer-made differentiation of images the phenomenon is a most undesirable one, since one of the elements differentiating pictures of various pathological lesions is, inter alia, the nature of edges of the objects analysed (the shape, sharpness of the separation, regularity, protrusion etc.). The object of the paper consists in working out an algorithm that would not cause distortion of the edges at the process of averaging.
JavaScript jest wyłączony w Twojej przeglądarce internetowej. Włącz go, a następnie odśwież stronę, aby móc w pełni z niej korzystać.