Design of a new Wavelet-based preprocessing stage for improved time-frequency detection of fetal movements Dr. Taoufik Ben Jabeur and Pr. Boualem Boashash Department of Electrical Engineering, Qatar University Background: Improving Newborn Health Outcomes can be achieved by monitoring Fetal Movement (FetMov) activity which relates directly to the fetus well-being. Several methods exist to record these Fetal movement (FetMov) signals including active (Ultrasounds) and passive ways (accelerometer). Ultrasound is an accurate measurement but expensive and intrusive; for these reasons, capturing FetMov through accelerometer sensors is a preferred option as it is low cost and nonintrusive. In this approach, different sensors are placed on the abdomen of the pregnant woman. Each sensor produces a FetMov signal. The recorded FetMov signals have been shown to be noisy and non-stationary signals containing several artifacts. Thus, a pre-processing stage is required to reduce the noise and eliminate the artifacts. Objectives: This work aims at designing a new pre-processing stage for improving FetMov detection by applying Wavelet de-noising algorithm followed by a threshold given by Kurtosis parameter. Methods: The new method is based on a Wavelet de-noising algorithm followed by a high pass filter to reduce some artifacts. A threshold given by Kurtosis parameter is used in the elimination of the artifacts. Results and conclusion: Simulation results show that the proposed pre-processing stage allows to reduce the noise and to eliminate more than 70% of the artifacts. Few artifacts are still present in the FetMov signals. As the FetMov signal is non-stationary, we use a Time Frequency Matched filter detector (TFMF) based on QTFDs to detect FetMov from these artifacts. The proposed preprocessing stage combined with TFMF detector improves significantly the performance of the detection of the FetMov signal leading to possible improvements in health outcomes for the fetus. Keywords: Fetal movement; Wavelet; Time frequency Matched Filter; Quadratic Time Frequency Distribution.


Article metrics loading...

Loading full text...

Full text loading...

This is a required field
Please enter a valid email address
Approval was a Success
Invalid data
An Error Occurred
Approval was partially successful, following selected items could not be processed due to error