Classification and suppression of noises in fetal heart rate monitoring: A survey

Abstract

Fetal heart rate (FHR) monitoring helps for detecting the fetal health status and intimates the decision for operative delivery to doctors at the earliest. Though it has got a greater significance, it poses extreme challenges in diagnosing the exact health condition because of noise intrusion either by internal or external sources. Irrespective of the sources, the noises can be classified based on their frequency of occurrence and amplitude. Hence, from the fetal heart rate information obtained from the abdomen of the mother, a simple SVM-based classification is done to distinguish the normal and the abnormal fetal heart rate. Later based on the type of noise interpreted, a suitable programmable filter technique is reviewed to remove unwanted information by varying the filter coefficients. A comprehensive set of features are chosen from MIT-BIH Arrhythmia Database. The analysis is carried over for each feature independent of the rest, and then it is generically continued by the automatic selection of features. The obtained results are classified based on similarities in features and spectrum. The simulations are performed using MATLAB and ModelSim. The area and timing analysis is evoked using Xilinx ISE.