Document Type : Original Article

Authors

• Ali Sheykhani ¹
• Farshad Hosseinzadeh Lotfi ²
• Arash Maghsoudi ¹

¹ Department of Biomedical Engineering, Science and Research Branch, Islamic Azad University, Tehran, Iran

² Department of Medical Engineering , Science and Research Branch, Islamic Azad University, Teran, Iran

Abstract

Worldwide, the rate of preterm births is increasing, so there will be significant health, development and economic problems. Premature birth is one of the leading causes of death and a significant cause for the loss of human potential among survivors around the world. Complications of preterm birth are the single largest direct cause of neonatal death. Current methods for early detection of such labor are insufficient. One promising technique, recognized in monitoring uterine activity, is the use of advanced device learning algorithms and electrohistrography (EHG) induction. In this article, a learning machine is designed to diagnose different types of deliveries. Using deep learning algorithms, electrohistrographic signals have been used to detect preterm birth. The results were obtained using a data set that included 262 cases for women who had a preterm delivery and 38 cases for women who had a preterm delivery. Using the "cross" technique, 4 types of data sets were implemented in two ways, with training and without training. The results obtained in this study showed that the error on this set of data was one percent.

Keywords

• Premature delivery
• Deep learning
• Electrohistrography
• Artificial intelligence

Main Subjects

• Optimization in science and engineering