The Development Of Perinatal Uterine Contraction Image Analysis System For Pregnant Women Based On Smart Phone

The division of perinatal period in China is from the 28th week of pregnancy to the 4th week after delivery.When pregnant women enter this period,there will be obvious and frequent uterine contractions.The intensity and frequency of uterine contractions are important indicators of clinical maternal and infant monitoring.At present,the fetal heart monitor is the most commonly used tool for clinical monitoring of pregnant women.The medical staff judges the health status of pregnant women and fetuses by observing the paper cardiotocogram.However,paper uterine contractions have shortcomings such as not easy to store,and subjective judgments of doctors are prone to deviations.So far,an effective tool has not yet appeared to be able to identify and determine the paper cardiotocogram.Therefore,it is of great significance to develop a monitoring system that can analyze and identify the paper cardiotocogram.This paper designs and develops a set of perinatal contraction image analysis system for pregnant women based on smart phones.First,using the mobile phone APP to call the mobile phone camera function to obtain the cardiotocogram;then,using the background program to perform pre-processing on the acquired image,extracting the contraction waveform,and taking the value of the contraction waveform data;and then analyzing and detenmining the extracted data;finally,presenting the results of the processing to the user through various functional modules of the APP,and uploads the contraction data and the judgment results to the cloud database.The main research contents of this paper are as follows:(1)Design and implementation of image preprocessing method for contraction waveform:According to the basic principles and methods of image processing,designed to achieve pre-processing of contraction images.The main steps include graying,denoising,rotation and correction of images and cropping of contraction images.(2)Design and implementation of image extraction method for contraction waveform:According to the operation principle and method of image morphology,designed to achieve the extraction of contractions.The main steps include image binarization,corrosion and expansion,connected domain noise reduction,clear the maximum contour and extract the data of the contraction waveform.(3)Design and implementation of uterine contraction waveform analysis and recognition algorithm:The quality of the contraction waveform is classified according to the different characteristics of the contraction waveform,determine the appropriate filtering method based on the waveform quality,and the baseline and normalization of the filtered waveform are performed.Finally,the health judgment result is given according to the criterion of effective uterine contraction.(4)Design and implementation of Android APP:Designed the overall architecture of the Android APP according to system requirements,and the principle description and effect display of each function module are carried out.The system realizes the mobile phone extraction,analysis and identification,health judgment,contraction waveform redraw,data upload and some basic functions.The system has completed the digital analysis and management of the paper cardiotocogram,overcomes the shortcomings of paper data that is difficult to preserve and easily damaged,and also provides real and reliable sample for future big data analysis research and artificial intelligence algorithm processing.