Quantitative Analysis Of DNA In Cervical Cells

We are in a period of rapid economic development.With the continuous improvement of the living standards,we pay more and more attention to our health.As a country with a high incidence of cervical cancer in the world,it's urgent to explore a large area screening.The traditional methods of screening for medical workers subject to subjective factors and the results of the diagnosis of the subjects are too professional,so that it's not easy to understand for a normal person without professional knowledge of medical pathology and it's necessary to develop a new method for early screening of cervical cancer with easier to extract,easier to operate,easier to read,higher automation,sensitiveness and specificity.In this paper,I have researched about DNA quantitative analysis based on the platform of automatic cervical cancer cell screening system,by which we can solve the problem that is automatically classifying the collected cells,and helping us to read the result report by machine.It realizes the automatic scanning reading,analysis and classification of cells with TBS cytology classification diagnosis and issue DNA quantitative analysis report and other functions,for the full implementation of the national cervical cancer screening provides a reliable and feasible technical means.Because of the nuclear overlap always happen by the methods of smear production in the process of cervical cell extraction,in this paper,we design a new model of active contour segmentation,by using GVF to structure energy function based on vector constraint to external forces,in order to get the contour curve evolution to the nucleus boundary constraint force of nuclear overlapping segmentation.The result of experiment show that it achieves the desired effect by using the method of overlapping nuclei segmentation,and it can be good on finding the concave point and determine the segmentation contour of the low overlapping cells image is preprocessed by morphological method can effectively reduce the interference of noise on image,and also it has faster speed.However,the highly overlapping cells can not achieve the desired segmentation effect.We studied the automatic classification of cervical cells with the morphological features,the optical density characteristics,the color features and texture features of the Gray-level Co-occurrence Matrix.The genetic algorithm based on the principle of "survival of the fittest"was used to pick up 34 kinds of all characteristic parameters which were favorable for cell classification and the greatest influence coefficient.The selection of these parameters including selection,exchange and mutation which contains the most information about the classification of cells.The experimental results show that the method is effective and reliable for the classification of impurities,epithelial cells,lymphocyte and inflammatory cells.Due to the DNA quantitative analysis of cervical cell only deals with epithelial cells,so that the classification method is reliable for the diagnosis results importantly.Because in the quantitative analysis of DNA in cervical cells,we calculate the DNA index only relate to epithelial cells and lymphocyte,so that a good cell classification of cells can reduce the wrong data about epithelial cell count,reduce the false positive rate effectively,improve the accuracy and reliability of the screening of cervical cancer cells.In this paper,we used the proposed method to detect the smears of the samples with different pathological changes by DNA quantitative analysis.Through the experiment,we verified that in the cervical cell samples,in the case of the confidence level of 0.95,the confidence interval of the DNA index of the cells in the nucleus was in a reasonable range of about 2,and the upper limit was less than 2.5,and also in the case of the confidence level of 0.95,the confidence interval of DNA index in the nucleus was within the reasonable range of the diploid cells.Diagnosis report generation including cell DNA index histogram,scatter diagram,part of the images and cervical cell ploidy information.It can make the patients reading the report under the guidance of a doctor clearly,and it's a good solution to the detection of cervical cancer cells by traditional method with more efficiency and provide a guarantee for the screening of large area.