Feature Extraction and Classification of VAG Signal of Human Knee for Early Diagnosis of Chondromalacia Patella

Abstract

Chondromalacia patella is a condition that is often left undiagnosed or misdiagnosed as osteoarthritis. Diagnosis of chondromalacia is difficult and requires CT scans for proper confirmation. Early detection of chondromalacia can prevent further damage to the cartilage. This paper aims at acquiring and processing different VAG signal from various individuals to ascertain if VAG signals can be used for the detection of knee pathologies and conditions like chondromalacia patella. In this paper, our primary discussion revolves around characterization of features of the signal from a piezosensor placed on the patella of the knee. There were marked differences observed between time domain correlation of the extracted characteristics of the VAG signals between subjects with knee pain, those with reported knee pathologies and healthy test subjects. VAG signals could thus be instrumental in early detection of chondromalacia patella and other knee pathologies.