An LPC-Based Approach to Heart Rhythm Estimation

Heart diseases are the cause for many of the human deaths in the world. A reasonable part of the population is affected by heart diseases. A primordial health monitoring to the diagnosis of heart illness is the electrocardiogram. The signal of the heartbeats captured in this protocol is key for medical professionals to investigate if the individuals are healthy or suffer from some possible heart disease. In this paper, the LPC technique is proposed to the automatic heart rhythm estimation. The QRS complex can have a variety of formats depending on the sensors positioning or even be hard to discriminate in a visual inspection of the signals or conventional threshold detection techniques. The signals are then filtered by a single, low order LPC filter, and the absolute value of the error estimation can be easily used in a threshold detection scheme to estimate the beats per minute heart rhythm. A database subset taken from MIT is used to test the efficiency of the approach. The direct measurement from the signals is presented to give a fair comparison of the findings.