The Research Of The Cervical Cells Satisfaction Automatic Evaluation Method

The early detection of cervical cancer is of great significance in reducingcervical cancer mortality.Liquid-based cervical cytology diagnosis level is in abackward state in our country,which specific performance in the drawn uncertaintiesdetermine the one-sidedness of the evidence, and rely on the experience of thediagnostician. And technical personnel only by the human eye cytologically, resultingin a large number of non-standard reports appear, even unmanned re-examination ofthe phenomenon, causing over treatment on clinical and medical accidents.It is very important for early cervical cell producer image. According to theBethesda2001meeting, smear test is divided into the "specimen assess satisfied" or"specimens assessment unsatisfactory" two categories. Proposed inflammation cellcoverage model, the specimens screened out whether satisfaction. According to theinterference factors of the inflammation cells, proposed satisfaction evaluation model,using the liquid-based cervical cells statistics and interpretation of their satisfaction onthe inflammation cell coverage model. The research focuses on the establishment ofinflammation cells coverage model to satisfaction rating, and by establishing thesatisfaction evaluation model to interpret the satisfaction, to achieve satisfactionevaluation faster image processing on the DSP development board. This thesis isapplied to the edge detection algorithm, removing impurities by morphological, colorspace segmentation method. For the research of the system, mainly to do the work ofthe following aspects:(1) The creation of inflammation cell coverage model:Through analysis the cervicalcell image, determined an effective segmentation method, established inflammationcell coverage model to achieve whether the liquid-based cervical cells is covered, caninitial realization for matlab simulation;(2)The creation of satisfaction evaluation model: To optimize the evaluation ofcervical cell image satisfaction, to be in-depth study of the satisfaction cervical cellsinfected extent, through the establishment of satisfaction evaluation model;(3) Taking into account the influence of impurities such as inflammation cells,byanalyzing the characteristics of the liquid-based cervical cells, proposed segmentationmethod based on the HSB color space and the impurity processing method based on morphology, solved the problem of interpretation of satisfaction evaluation. Finally,the use of fault-tolerant technology to optimize the satisfaction evaluation system.(4) Through the theory that mentioned, the author applied the satisfaction evaluationsystem to the practice, can achieve portable function in the future.The paper makes an in-depth research to the key technologies of the system fromboth theoretical and applied aspects, the research results show that dealing with theimage of a liquid-based cervical cell producer satisfaction time less than500ms, andthe accuracy rate of more than90%. The application of this method can achieve goodresults in the rapid evaluation of liquid-based cervical cell producer satisfaction, it isfeasible in engineering applications.