A press conference was conducted jointly by Indian Institute of Technology Kharagpur (IIT Kharagpur) and Institute of Pulmocare & Research Kolkata on July 09, 2015. This was related to peer recognition of novel findings of a collaborative research, taken up by two institutes (Respirology: Editorial, Editor’s Choice). Director, IIT Kharagpur in his public message congratulated the researchers from either side for “achieving this milestone in our strategic project initiative ‘Signals & Systems for Life Science’.” The following had been the press release on this occasion. The work drew attention of a large section of media that includes Times of India, Business Standard, Economic Times, Indian Express, Hindustan Times, The Statesman, Echo of India, Sify News, Zee News.
Now Doctors can Observe Sound patterns to Detect Lung Disorders
July 9, 2015, Kolkata: “Seeing what we hear: an eye to help the ear”: This was the title of a recent editorial in the reputed medical journal ‘Respirology’ published in May, 2015. The entire editorial focused on the new findings of a collaborative research between Indian Institute of Technology Kharagpur (IITKGP) and Institute of Pulmocare & Research Kolkata (IPCR). The work was related to seeing something that we hear, rather cannot hear well enough!
Human listenting ability, that can distinguish the voice of a crow from the voice of a cuckoo, goes beyond appreciation of pleasant and unpleasant sound in our external environment. The invention of stethoscope in 1816 showed its great potential in medical field. Auscultation, i.e. listening to internal sounds of the body, usually through stethoscope, became an important component of clinical examination. Physicians through training started developing specific skills to distinguish normal and abnormal body generated sounds, specially that of heart and lung sounds. However, there is an element of subjectivity in appreciation of stethoscope sound due to variation and deterioation of individual’s listening ability; also the expertize required to distinguish such sounds that come through professional training may not be available everywhere, more particularly in our rural backyard.
The researchers at IPCR and IITKGP through a collaboration started investigating how the correlation between lung mechanics (normal or abnormal) and acoustic indices could be put in an alternate form of representation that is easier to comprehend, and would be useful for diagnostic purpose. The target was to identify certain patterns in the sound which carry distinguishing signatures of normality or abnormality, and present it in a visual form for easier identification of the same. For this, lung sound was collected from two groups of people One group belonged to normal and the other group had a specific lung disease. Advanced digital signal processing and feature extraction algorithms were used to identify distinctive characteristics of these two kind of signals. A mapping algorithm was developed to convert the digitally processed lung sound into pictures that highlight these distinguishing features. It was found that in spite of subject to subject variability in data collection, these pictures were representative enough to separate normality from abnormality, and the same was verified by three independent physicians. The exercise was extended to generate an intestering finding. Three independent laymen when presented with these pictures, could also differentiate one kind of signal from the other.
The researchers feel that the work has the potential in filling up the much felt gap of lack of quality healthcare facilities in remote places. It can also serve as an objective assisting tool to all physicians to help them in diagnosis. The work ahead is to increase the number of disease portfolio, accordingly making improvement in the technology, validating the result through independent observers, improving acceptability for adoption of it in clinical practice through more scientific investigations.
Editors of the ‘Respirology’ journal in explaining why they are excited about this work and why it became the topic of the editorial said the following: “The study is novel in more than one way. It is novel in the aspects of sound that were chosen for analysis. It is novel (at least for us as physicians) in its visual representation of sound. The struggle to convert sound into images could be considered as herculean as the struggle to convert language into a written form, and the authors have presented an interpretable pictorial display of their data.”
Three videos related to this press conference can be found at