Hybrid Algorithm To Classify The Lung Abnormalities Using The Stethoscope Recordings

Abstract

The pandemic of 2020 brought a lot of changes to the health and medical industry where a lot of smart devices started flowing in to compensate the lack of hospitals for the less severe cases. This work primarily focuses on the lung abnormalities as the lungs were one of the first organs to break down when the virus affected the body. The lung abnormalities whilst not becoming severe but will act as a catalyst to spread the virus’s effect through the lung thus concluding in complete pulmonary breakdown or irreversible lung damage. This work focuses on hybridisation of Fast Independent Component Analysis (FAST-ICA) algorithm coupled with Mel-frequency cepstral coefficients (MFCC) for independent component analysis and is fed as a input to the classification algorithm of Keras library which is the sequential algorithm. The pre-processing of the data in FAST-ICA is mainly based on three components of sound which are frequency, amplitude and wavelength.