A novel algorithm for spirometric signal processing and classification by evolutionary approach and its implementation on an ARM embedded platform

Spirometry is the most commonly performed Pulmonary Function Test (PFT) which is used to distinguish obstructive from restrictive lung diseases. This paper presents the basic system requirements for an automatic pulmonary disease classification system based on spirometric signal using a novel algorithm. The software of the system extracted features from the digitized spirogram waveform values and classified the disorders with minimum uncertainty. Classification was done by generating more data from the available trials/ tests using an Evolutionary Approach called Genetic Algorithm (GA) and without using any prediction equations as done by the conventional spirometers. Thus GA ensures reduction in the number of trials to be performed by the patient there by reducing patient stress. The hardware requirements for implementation on an embedded system are also presented. On an average, the accuracy was found to be 95.74%.