PROJECT TITLE :
Wavelet Augmented Cough Analysis for Rapid Childhood Pneumonia Diagnosis
Pneumonia is the cause of death for over 1,000,000 youngsters every year around the world, largely in resource poor regions such as sub-Saharan Africa and remote Asia. One of the largest challenges faced by pneumonia endemic countries is that the absence of a field deployable diagnostic tool that's rapid, low-cost and accurate. In this paper, we have a tendency to address this issue and propose a methodology to screen pneumonia primarily based on the mathematical analysis of cough sounds. In explicit, we propose a unique cough feature impressed by wavelet-primarily based crackle detection work in lung sound analysis. These options are then combined with different mathematical features to develop an automated machine classifier, which will separate pneumonia from a vary of alternative respiratory diseases. Both cough and crackles are symptoms of pneumonia, however their existence alone is not a specific enough marker of the disease. In this paper, we tend to hypothesize that the mathematical analysis of cough sounds permits us to diagnose pneumonia with sufficient sensitivity and specificity. Using a bedside microphone, we tend to collected 815 cough sounds from 91 patients with respiratory illnesses like pneumonia, asthma, and bronchitis. We extracted wavelet features from cough sounds and combined them with different features like Mel Cepstral coefficients and non-Gaussianity index. We then trained a logistic regression classifier to separate pneumonia from different diseases. As the reference commonplace, we used the diagnosis by physicians aided with laboratory and radiological results as deemed necessary for a clinical call. The strategies proposed in this paper achieved a sensitivity and specificity of ninety four% and 63%, respectively, in separating pneumonia patients from non-pneumonia patients based mostly on wavelet options alone. Combining the wavelets with options from our previous work improves the performance any to 94% and eighty eight% sensitivity and specificity. The performance so much surpasses that of the WHO criteria cu- rently in common use in resource-limited settings.
Did you like this research project?
To get this research project Guidelines, Training and Code... Click Here