Study For Measurement Method Of Fetal Head Position By Combining Attitude Detection And Ultrasound Imaging

Fetal head position(FHP)is one of the important indicators of childbirth.Abnormal head position is likely to affect the normal course of childbirth,which may lead to increased risks and costs of childbirth.The monitoring of FHP is indispensable for the reasonable assessment of the progress of delivery.The risk of childbirth can be reduced by the timely detection of abnormal FHP and the initiation of intervention measures.Vaginal examination as a invasive methods is commonly used in clinical practice,which is also highly subjective and easily increase the risk of maternal discomfort and infection.Ultrasound examination,as a non-invasive method of examination,is more easily accepted by mothers.And with the help of ultrasonic equipment to evaluate the FHP,the evaluation results can be quantified,which is convenient for technology transmission.However,the use of ultrasound equipment to assess FHP requires the operator to have proficient ultrasound skills,extensive obstetric experience,and good spatial imagery translation.Assessment results are prone to vary widely depending on the operator’s experience.Due to the barrier of experience threshold,it is difficult for doctors and midwives to evaluate the FHP with ultrasonic equipment,and the technology is difficult to be popularized.Aiming at the above problems,this paper proposes a new method for FHP measurement by fusion attitude detection and ultrasonic imaging.The key techniques of attitude detection and the method of fetel head position measurement by combining on nine aixs sensor and ultraound image are studied and verified by experiments.Firstly,the relationship between the ultrasound image of abdomen scan and the attitude of probe is analyzed.A method for determining FHP based on probe attitude and anatomical features in ultrasound images is proposed and a technical scheme is designed.Based on the characteristics of fetal head ultrasound images,an optimized network structure is proposed in combination with U-net and VGG9 deep learning network to achieve automatic judgment of FHP and improve efficiency.Secondly,an experimental platform for attitude detection was designed,and Mahony algorithm was used for attitude calculation and simulation analysis.Then,the threshold detection methods are used to improve the attitude solving algorithm,so that the attitude solving algorithm can meet the requirements of the interference environment of the hospital departments.Finally,the accuracy and robustness of the attitude solution are verified,and the efficiency and accuracy of the deep learning network are verified.Consistency analysis is performed to compare the difference between the results of the FHP calculation and the benchmark value.The experimental results show that the output of the calibrated attitude detection module has a good linearity,the error of the heading angle measurement is less than2°under the interference-free condition.An anti-magnetic interference analysis of the fusion algorithm showed that the heading angle error could be suppressed to within 3°by setting a suitable threshold,which provides the basis for measuring the probe deflection angle in the measurement method used in this paper.The average error of the deep learning network used to identify the FHP is about 2.2°,with a classification accuracy of 96.2%.The Bland-Altman concordance analysis obtained from the FHP simulations measurement showed good agreement with the baseline values of the model,with S_d and RMSE of 0.96°and 1.7°,respectively.Hence,this paper proposed a method that can accurately measure the FHP,provided a reliable method for assisting doctors to monitor the progress of labor,reduce the threshold of experience needed by doctors to evaluate the FHP,and help doctors to timely detect abnormalities in labor and initiate scientific intervention means.