Automated pediatric cardiac auscultation

[摘要] Most of the relevant and severe congenital cardiac malfunctions can be recognizedin the neonatal period of a child's life. The delayed recognition of a congenital heartdefect may have a serious impact on the long-term outcome of the affected child.Experienced cardiologists can usually evaluate heart murmurs with a high sensitivityand specificity, although non-specialists, with less clinical experience, may havemore difficulty. Although primary care physicians frequently encounter childrenwith heart murmurs most of these murmurs are innocent.The aim of this project is to design an automated algorithm that can assist the primarycare physician in screening and diagnosing pediatric patients with possiblecardiac malfunctions. Although attempts have been made to automate screening byauscultation, no device is currently available to fulfill this function. Multiple indicatorsof pathology are nonetheless available from heart sounds and were elicitedusing several signal processing techniques. The three feature extraction algorithms(FEA's) developed respectively made use of a Direct Ratio technique, a Waveletanalysis technique and a Knowledge based neural network technique. Several implementationsof each technique are evaluated to identify the best performer. Totest the performance of the various algorithms, the clinical auscultation sounds andECG-data of 163 patients, aged between 2 months and 16 years, were digitized.Results presented show that the De-noised Jack-Knife neural network can classify 163recordings with a sensitivity and specificity of 92 % and 92.9 % respectively. Thisstudy concludes that, in certain conditions, the developed automated auscultationalgorithms show significant potential in their use as an alternative evaluation techniquefor the classification of heart sounds in normal (innocent) and pathologicalclasses.

[发布机构] Stellenbosch University