federalnoe gosudarstvennoe avtonomnoe obrazovatelnoe uchrezhdenie vysshego obrazovaniya "Sankt-Peterburgskij politekhnicheskij universitet Petra Velikogo" (FGAOU VO "SPbPU")
发明人:
Utkin Lev Vladimirovich (RU),Уткин Лев Владимирович (RU),Meldo Anna Aleksandrovna (RU),Мелдо Анна Александровна (RU),Ryabinin Mikhail Andreevich (RU),Рябинин Михаил Андреевич (RU),Lukashin Aleksej And
申请号：
RU2018141208
公开号：
RU0002694476C1
申请日:
2018.11.22
申请国别(地区):
RU
年份:
2019
代理人:
摘要：
FIELD: medicine.SUBSTANCE: invention refers to medicine and aims at intelligent diagnosis of lung cancer. Disclosed is a method for detecting and diagnosing lung cancer based on intelligent analysis of the shape, malignant neoplasm structures in the lungs, involving treatment of patient's lung images, obtained by computed tomography, as a result of which a graphic image is masked voxels with Densitometric density values according to Hounsfield scale with not corresponding to density values of lung tissues, following segmentation of voxels located on the surface and inside the "candidates" of new growths, constructing "inner" chords formed by combinations of pairs of points located in allocated voxels on the surface of "candidates" of new growths, growth of histogram of distribution of lengths of "internal" chords for each "candidate" with reduction to maximum length of "internal" chord constructed within borders of each "candidate" of new growth, constructing a histogram of distribution of Densitometric Density on the Hounsfield scale inside each candidate of the new growth with bringing to the maximum value of Densitometric Density on the Hounsfield scale, defined in random points on the "internal" chords, constructing "external" chords formed by combinations of pairs of points located on the surface of the "candidate" of the growth and on the faces of the cube, constructed around the "candidate" of the growth, constructing for each "candidate" of the growth of the histogram of distribution of lengths of "external" chords with reduction to the maximum length of the "external" chord, constructing a histogram of distribution of Densitometric Density on the Hounsfield scale inside each candidate of the new growth with bringing to the maximum value of Densitometric density on the Hounsfield scale, determined in random points on the "external" chords, forming a feature vector comprising data of four plotted histograms, followed by classification of each "candidate" of n
