文献阅读4

最新推荐文章于 2024-10-04 20:50:17 发布

Erwachen

最新推荐文章于 2024-10-04 20:50:17 发布

阅读量95

点赞数

分类专栏：文献整理(个人使用) 文章标签：深度学习人工智能

本文链接：https://blog.csdn.net/Erwachen/article/details/128374058

版权

文献整理(个人使用) 专栏收录该内容

8 篇文章 1 订阅

订阅专栏

该研究提出了一种利用Morse小波提取胎心率（FHR）信号的时间频率信息，并通过预训练的ResNet50模型进行细调，以检测胎儿窘迫的深度学习方法。在第一产程的FHR数据上，实现了98.7%的准确性、97.0%的敏感性和100%的特异性；在第二产程的数据上，准确率为96.1%，敏感性为94.1%，特异性为97.7%。该模型可以作为妇产科医生的决策辅助工具。

摘要由CSDN通过智能技术生成

06

Date:2022.12.19--06
Title:Deep learning based fetal distress detection from time frequency representation of cardiotocogram signal using Morse wavelet: research study Link:Deep learning based fetal distress detection from time frequency representation of cardiotocogram signal using Morse wavelet: research study - PMC (nih.gov)D
Dataset	CTU-UHB: only FHR From the CTG data that is obtained from the only open access CTU-UHB data base only FHR signal is extracted and preprocessed to remove noises and spikes.
Framework
Methodology	After preprocessing the time frequency information of FHR signal is extracted by using generalized Morse wavelet and fed to a pre-trained ResNet 50 model which is fine tuned and configured according to the dataset.
Results	A promising classification result which is accuracy of 98.7%, sensitivity of 97.0% and specificity 100% are achieved for FHR signal of 1st stage of labor. For FHR recorded in last stage of labor, accuracy of 96.1%, sensitivity of 94.1% and specificity 97.7% are achieved.
Contributions	The key contribution of the current study is as follows; preprocessing done in the first step on the raw FHR signal was to remove unwanted artifacts and missing data.Converting the 1D FHR signal to 2D image by applying effective method of time–frequency representation was done in the second step. Fine tuning of the ResNet 50 deep learning model for training and classification was done in the third step. The last step was used to conduct comprehensive experiment on FHR signal that were recorded on early (first stage) and last (second) stage of labor. Since the quality and nature of FHR signal varies depending on the labor stage when it was recorded [23], a comprehensive experiment has been done for FHR signal recorded in varies stage of labor to confirm robustness of the mode.
Conclusion	This study proposed an automatic system for fetal distress detection from time frequency information of CTG signal using generalized Morse wavelet. A pre-trained ResNet 50 model was used to classify fetal condition as normal and distressed for first stage and last stage of labor FHR data. The developed model can be used as a decision-making aid system for obstetrician and gynecologist.
Notes	Siganal process and it's writing.