Research On Intelligent Classification Technology Of Cervical Cancer Cell Microscopic Image

Cervical cancer is one of the most common cancers that threaten women’s health,with more than 500,000 new cases every year around the world.Among them,developing countries have always had a high incidence due to limited medical conditions.my country is the most populous country in the world and the largest developing country.The incidence of cervical cancer ranks first in the world every year.The increase in the number of cases accounted for about 30% of the world’s total,seriously endangering the lives of women in China,and it is urgent to curb the spread of cervical cancer in my country.In the detection methods of cervical cancer,the microscopic image examination of cervical cells is recognized as the simplest,most economical and most direct method.This method is to observe the microscopic images of cervical cells,and doctors can effectively judge whether the cervical cells are pathological based on professional medical knowledge,which has a huge effect on the mortality of cervical cancer.However,observing the microscopic images of cervical cells requires professional doctors.Our country has a vast territory,a large population,and the distribution of medical resources is extremely unbalanced.For areas with relatively backward medical conditions,there are very few professional medical personnel,and it is impossible to complete the cervix.Cancer screening work.At the same time,for a doctor who specializes in cell testing,because everyone’s energy is limited,the number of cervical cells checked every day will be greatly limited.For the huge number of cervical cancer incidence in my country,the detection efficiency is completely Unable to meet the requirements for cervical cancer screening.At the same time,because doctors who observe cell images work for a long time,they will inevitably have fatigue,and the correct detection rate of cervical cells will be greatly reduced,which is not conducive to the correct screening of cervical cells.Based on this,researching intelligent classification technology to complete the classification of cervical pathological cells can detect and treat cervical cancer more quickly.This paper mainly studies the intelligent classification technology of cervical cancer cells,including cervical cell preprocessing,image segmentation,feature extraction,detection and recognition.The main research work of this paper includes:First,the research status and significance of intelligent classification of cervical cancer cells.Second,to study the related theories of cervical cancer cell microscopic images and collect cervical cell image data sets.Third,preprocess the image to improve the quality of the image,which includes operations such as grayscale,denoising,and image enhancement.Compare the segmentation effect of the traditional image segmentation algorithm with the largest interclass algorithm and the watershed algorithm.The U-net network based on the deep learning network is introduced.The 7000 images are first marked as a data set at the pixel level,and finally the cell is processed by the improved U-net network.Split.Fourth,for image classification,the traditional feature extraction method is compared with the feature extraction method based on deep learning,the feature extraction method based on transfer learning is selected,and the traditional classification algorithm is compared with the deep learning neural network.The best classification algorithm is ResNet,which has an accuracy rate of 94.31%,a recall rate of 90.76%,an f1 value of 0.9243,and a kappa coefficient of 0.9143.Fifth,for the practical application of the intelligent classification technology of cervical cancer cell microscopic images,according to the MVC design model,the QT development framework and My SQL database are used to implement intelligent software that integrates detection-related algorithms.