HSIL Detection Of Colposcopy Based On Deep Learning

Cervical cancer is one of the most common cancers in the world,and is the most easily treated cancer at the early stage of discovery.Cervical cancer is not accompanied by obvious clinical symptoms before or after onset.Therefore,cervical cancer detection can only be achieved through high cost screening.Because of the high cost,cervical cancer screening can't be popularized in some developing countries and underdeveloped areas,which makes the mortality of cervical cancer in these areas greatly increased.This article first introduces the research status of cervical cancer detection at home and abroad,and the significance of the image analysis of colposcopy for cervical cancer screening.Then a colposcopy image analysis method based on depth learning was proposed to detect cervical cancer and precancerous lesions at a lower cost.At the same time,in order to make this method more popularized,a location method of lesion area is also proposed,which is combined with the test results to achieve qualitative and quantitative analysis in all aspects.At the end of the article,the final discrimination effect of the method and the location effect of the focus area are given through the comparison of various experiments.