Research On Non-destructive Detection Method Of Uterine Cold Infertility Based On Infrared Thermal Imaging Technology

Infertility has a great negative impact on the body and mind of patients and their families.Conventional detection methods such as hydrotubation,hysterosalpingography or hysteroscopy-laparoscopy have the risk of minimal invasion and infection,and can not detect the uterine cold infertility.therefore,it is of great significance to explore a set of low-cost,accurate and non-destructive detection method of uterine cold infertility.In this paper,a method combining infrared thermal imaging technology and deep learning is proposed to detect uterine cold infertility.The infrared images of the abdomen of women of childbearing age were collected to compare the temperature distribution of the abdomen of patients with uterine cold infertility and infertile women of childbearing age,and to explore the relationship between abnormal temperature distribution and uterine cold infertility.SVM,vgg16 model and Inception V3 model were used to model and recognize the infrared image of uterine cold infertility respectively.The advantages and disadvantages of the three recognition methods were compared,and the optimal recognition method was selected to establish the auxiliary monitoring and diagnosis system of uterine cold infertility.The results showed that the abdominal temperature distribution of the patients with uterine cold infertility was not uniform,and there were obvious cool spots in the hypogastrium,and the temperature was lower than that of infertile women of childbearing age.The difference of abdominal surface temperature of patients with uterine cold type infertility was greater.The average difference of temperature difference was tested by Student's t test,and the difference was statistically significant(P(27)0.01).The uterine cold infertility was related to the abnormal distribution of human body temperature.SVM,VGG16 and Inception V3 were used to model and recognize the infrared image of abdomen,and the accuracy rates were 95 %,98.75 %and 95.25 % respectively.The recognition technology based on infrared thermal imaging and machine vision can effectively detect and identify the uterine cold infertility.This study is to analyze the images collected in the laboratory under ideal conditions.In the actual medical detection,it is impossible to eliminate the influence of interference factors.Therefore,this study can not be used as an independent detection method of uterine cold infertility.In the future,we should combine different interference infrared thermography for modeling and identification,and establish a more universal detection method for uterine cold infertility.