Researchers and faculty at the Dept. of Electronics and Electrical Communication, IIT Kharagpur have come up with a heart sound analyzer to facilitate detection of valvular heart diseases from heart sound. This is an integrated device for signal conditioning, data acquisition, analysis, characterization and automated decision support of heart sound signals. This low cost, noninvasive, primary diagnostic device can be used by both experts and non-experts and is directed towards making healthcare more efficient and affordable.
The data acquisition subsystem involves a low cost hardware connected to a simple stethoscope that does appropriate signal conditioning for digital data capture. The data compression and decompression subsystem is used for compressing and encrypting the data in a novel .hsa format to reduce storage space and maintain privacy of data. The recording and display subsystem records the data in real time and also handles online/offline displays, zooming, playback of selected part, playback in loop, patient database management, report generation and printing. The decision making subsystem characterizes the recorded data online/offline in manual mode for thirty three different pathological cases through visual comparison of features or audio clues and in auto assistant mode for nine commonly occurring diseases by a system generated suggestion of probable disease, if any, with a confidence score. This can also make feature wise comparison of patient’s previous record with current one helping analyze the progression of disease. The heart sound analyzer can also be used as an auscultation training device for a medical professional.
The heart and lung sound analyzer was recently exhibited at Rashtrapati Bhavan in an event organized by Indian Council of Medical Research (ICMR) with the theme “Innovation in Medical Science & Technology” as a part of week-long “Festival of innovation” that was inaugurated by Honourable President of India
The heart sound analyzer is a low-cost, portable, user-friendly, primary diagnostic tool. It will prove to be very useful in bringing people closer to healthcare professionals and in turn ensure quality health service at affordable cost. Some multinational companies have already shown interest in funding further research and marketing the product. The product was awarded DST Lockheed Martin India Innovation Growth Program and has already been filed for patent. It has been successfully displayed at events in Pragati Maidan, Delhi and Taj GVK, Hyderabad.
Dr. Samit Ari, Dr.Suman Senapati, and research students A. Kishore Kumar, Ashok Mandal, at the Audio and Bio-Signal Processing Lab., Dept. of Electronics & ECE, IIT Kharagpur have carried on this research under the able guidance of Dr. Goutam Saha., Associate Professor, Dept. of Electronics and Electrical Communication, IIT Kharagpur. Email: gsaha@ece.iitkgp.ernet.in
A featured video of ‘Heart Sound Analyzer’ appears in https://www.indiainnovates.in/technologies.aspx filmed by FICCI few months ago.
In another research project Dr. Goutam Saha, along with his past and current students Dr. Sandipan Chakraborty, MD. Sahidullah, Neeraj Bhende, Israj Ali, Prashanta Pattadar, Susanta Sarangi, Monisankha Pal, Debapriya Sengupta have come up with a voice based authentication device. It uses voice of a person to verify his / her identity over a telephone line. A large number of services today are voice based and uses IVR in interacting with a customer. This device can be used to assess identity of customer and services, eligible to such customers, can be offered.
The voice data is collected through a voice modem connected to a computer using Telephony Application Programming Interface. An IVR system developed for this guides the speaker in the entire process. Registration of voice samples takes place through a controlled environment. The verification compares voice samples with registration data in a pattern recognition framework. A feature extractor extracts relevant features and then develops a speaker model which remains stored in the computer. The voice to be verified passes through this speaker model and accept or reject decision is taken based on score generated. The prototype developed is a proof of concept for real-time voice based biometric authentication of a person.
The device has been demonstrated at Space Application Centre (SAC), ISRO, Ahmedabad. It has also been successfully demonstrated speech based prototype that identifies a speaker over VoIP platform, suitable for skype like conversation, to our another sponsor, Vodafone and this one was hailed among the ‘Top 100’ innovations by in India Innovation Growth Program 2013 conducted by DST, Lockheed Martin, FICCI, Indo-US S & T Forum, Stanford Business Graduate School and IC2 Institute.
