Tuberculosis Type Iii Lesions Aided Diagnosis Positioning System

The medical picture is the important objective basis diagnosed that the doctor carries on clinical diagnosis, following state of an illness, operation plan, prognosis studying, antidiastole. The X-rays imaging technology has many characteristic, such as simple ,a little painful, cheap , the high space resolution etc., still widely used in each hospital now, and check the project as the first-selection of the skeleton system , respiratory system ,etc, especially have a very important position in diagnosis of the pulmonary tuberculosis.With the characteristic of image of the Model III pulmonary tuberculosis X-rays chest, I studydeeply about localization and drawing effectively fast. The paper divides four parts mainly. (1)Preprocess of the X-rays chest image: The ones mainly include that the filtering processing toprimitive image and binarization processing using edge detection method based on Cannyoperator .We get some parameters of the Canny operator with analyzing and experiment andthese parameters are good at the medical image,such as low threshold value / high thresholdvalue, high threshold value / the count of pixel of image and the standard square estimation of theGauss filter; (2) Testing the line of upper and lower edge of rib: Utilizing the medical anatomyscience knowledge, this paper has put forward the improving direction operator(the direction ofleft rib and right rib),which includes left direction operator and right direction operator, and utilize itto follow and draw the picture with line of upper and lower edge of rib to the binary image ofoperated based on Canny; (3) Jointing and fill-up the rib line which has disconnection and lackcharacteristic and the fitting parabolas : According to the rib line information followed, the paperhas put forward a kind of effective fast algorithm about adjoining and mending the rib lines,and themain thinking is that we propose the method of the high weight searching. The method will scaneach line in reserve, and then get some a collection of candidate points to each point. We willselect a pair of point, which has the most high weigh value, to deal with the picture using LinearInterpolation; at last joins it as a rib line, and then fitting all rib line image; (4) The focus makes areservation: The text has proposed an algorithm of focus location which accords with the rule ofthe clinical medicine naming, and this algorithm utilizes the parabola fitting in batches to locateaccurately fast.The research of this thesis has utilized technology of the computer and image processing technology to test and draw out rib and focus in the chest image automatically mainly, having facilitated the analysis on a later stage of chest and drawn the relevant data, and it has important meanings for liberating manpower, saves time, reducing the misdiagnosing rate.