Detection of cardiac arrhythmic beats by logical classifier using binary coding

Cardiac arrhythmia is a rhythmic disorder in heart generated from problems in impulse formation or conduction. Automatic detection of arrhythmia at the time of its occurrence is very essential for continuous cardiac monitoring. As different kinds of arrhythmias produce different morphology of beats, it is difficult to suggest one typical algorithm to classify all kind of arrhythmic beats. In this work, a binary logical classifier is proposed to identify beats such as premature ventricular contraction, bundle branch block, paced and normal. In most of the arrhythmic beats the wave texture is modified along with a change in temporal information. The proposed logical classification requires binary coding of beats based on temporal and morphological features to form a binary array representing the status or signature of the beat. Next the classification is made by the status of the bits depending upon its position in the binary array. This approach of logical classification reduces the complexity of detection because of its binary nature resulting in fast and ready reference of arrhythmic disorder for standalone systems. Performance evaluation results indicate that the method has good accuracy, sensitivity and positive predictivity for all kind of test beats. Moreover, this technique is easily hardware implementable because of its binary nature.