Automated Analysis of Lung Sounds as a Predictor of Ventilator Associated Pneumonia (VAP)

Lead Research Organisation: University of Southampton
Department Name: Faculty of Engineering & the Environment


This project will make use of breath sounds, recorded using an electronic stethoscope, to detect lung problems in patients on intensive care units. Listening to the lungs with a traditional stethoscope provides an assessment of airway and lung tissue health, but the information it gives is subjective and qualitative. Computer aided lung sound analysis (CALSA) uses a digital stethoscope to record sounds for processing, removing the subjectivity and allowing quantification of their acoustic characteristics. Information derived from CALSA has the potential to aid the early detection, diagnosis and monitoring of lung complications, which are common in post-operative and intensive care patients. Mechanically ventilated patients are at high risk for developing ventilator-associated pneumonia (VAP), which is a significant cause of morbidity and mortality; one symptom of this is a change to the lung sounds.

In this study the focus will be on signal processing and analysis of lung sound recordings to evaluate the usefulness of extracted acoustics features for predicting and monitoring development of VAP. Development of signal conditioning techniques to remove unwanted sounds will be followed by application of feature extraction and selection techniques to identify the subset of acoustic features most sensitively predictive of VAP onset and progression. The goal is a robust and reliable system that can easily be used by health service professionals in clinical surroundings.



Ravi Pal (Student)


10 25 50
publication icon
Pal R (2019) A dataset for systematic testing of crackle separation techniques. in Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Annual International Conference

Studentship Projects

Project Reference Relationship Related To Start End Student Name
EP/N509747/1 01/10/2016 30/09/2021
1992735 Studentship EP/N509747/1 01/11/2017 31/10/2020 Ravi Pal
Description Pulmonary crackles are indicative of lung pathology and may be used for diagnosis and monitoring of disease. Many algorithms have been proposed to separate the crackle sounds from the breath noise, but a lack of standardized processes for evaluating their performance makes comparisons difficult. So we have proposed a standard data set to be used for systematic comparative testing.
Exploitation Route The proposed data set can provide a useful comparative performance analysis of different crackle separation algorithms. Researchers can use this as a standard data set for testing separation algorithms which is available at
Sectors Healthcare

Title A dataset for systematic testing of crackle separation techniques 
Description This dataset contains: (a) real fine and coarse crackles, and simulated fine and coarse crackles with different initial deflection width and two-cycle deflection width, (b) Matlab code for generating noises with different spectral properties and different SNR (range -10dB to 10dB) compared to the crackles: i) noise with the spectral characteristics of healthy breath noise and ii) Gaussian white noise and (c) real lung sound samples with i) predominantly fine inspiratory crackles (idiopathic pulmonary fibrosis) and ii) predominantly coarse crackles in both inspiration and expiration (bronchiectasis). 
Type Of Material Database/Collection of data 
Year Produced 2019 
Provided To Others? Yes  
Impact Researchers can use this dataset for systematic testing of crackle separation techniques.