Tidc2007 cardionet
•Download as PPT, PDF•
0 likes•314 views
This document summarizes a presentation on developing novel healthcare solutions using multimedia, wireless, and sensor technologies. It discusses the vision of making cardiac healthcare more patient-centric by applying advances in these areas. The key components discussed are a wearable 1/3/12-lead ECG recorder, a 12-lead ECG recorder, a remote video consultation system, and a cardiac decision support system. Use case scenarios involving patients at home and in ambulances are presented. The implementation and future work focus on developing an intelligent cardiac network for real-time remote doctor consultations using low-cost, low-bandwidth solutions.
More Related Content
What's hot
What's hot (20)
Viewers also liked
Viewers also liked (20)
Similar to Tidc2007 cardionet
Similar to Tidc2007 cardionet (20)
More from Arpan Pal
More from Arpan Pal (20)
Tidc2007 cardionet
- 1. Dr. B.S.Adiga, Arpan Pal Tata Consultancy Services India Using Multimedia, Wireless and Sensor Technology to Develop Novel Healthcare Solutions
- 2. 2 Presentation Roadmap Introduction Use Case Scenarios Components Implementation and Demo
- 3. 3 Vision and Mission • To make Cardiac Health Care more patient centric than hospital centric • To apply advances in the areas of Multimedia, Mobile Technology and microelectronics to bring expert cardiac health care to rural masses. • To design and develop a prototype of TCS Cardionet as proof of concept • To conduct field trials at various hospitals in the country and arrive at comprehensive product specifications
- 4. 4 Motivation • Lack of specialized cardiologists at primary health centers • Need of a low cost solution for the rural masses • Need for constant cardiac monitoring
- 5. 5 The Solution • A low cost secondary e-consultation service • Augmented by a centralized Decision Support System (DSS) • Assists medical practitioners to expedite interpretation and diagnosis of cardiac ailments • Intended to provide life saving gadgets at low cost • Aims at providing low cost telecardiology services • To the rural masses of the country at their door steps • To the ambulatory patients
- 6. 6 Presentation Roadmap Introduction Use Case Scenarios Components Implementation and Demo
- 7. 7 Intelligent Cardionet Patient at Home Patient in Ambulance Hospital There is need for collaboration with doctor • In real-time • In a cost-effective manner • Using low-bandwidth
- 8. 8 Patient @ Home Hospital Patient at Home
- 9. 9 Patient @ Ambulance Audio/Video and ECG data from ambulance Patient in Ambulance Hospital Audio data from doctor Doctor’s View at his laptop/desktop
- 10. 10 Presentation Roadmap Introduction Use Case Scenarios Components Implementation and Demo
- 11. 11 Components 1/3/12 Lead Wearable ECG Recorder 12 Lead, ECG Recorder ECG Decision Support SystemRemote Video Consultation System
- 12. 12 1/3/12 Lead Wearable ECG Recorder • MSP430F149: 16 bit Ultra low- power microcontroller from TI • Sampling rate of 360 samples per second • 12 bit A/D Converter • 32 MB Flash for data storage • 3.7V mobile battery • Low Power • Low-cost • Lightweight
- 13. 13 12 Lead Recorder • 12 Lead ECG System • Leads: I, II, III, aVR, aVL, aVF, V • Standard Calibration pulse • Gain adjustment in ½ and 1 gain modes • Acquisition of ECG @ 256 samples per second • Thermal Printer for printing ECG on graph (at std rates of 25mm/sec, 50mm/sec) • Battery powered and Internal battery charging • Portable
- 14. 14 Remote Video Consultation System • H.264 based video, AMR based audio • Constant bit-rate (CBR) and Variable bit-rate (VBR) support • Low-bandwidth / High-quality • Proprietary error resilience algorithms for robust performance • One-way or two-way video • Instrumentation Signals / any other application data can be shared remotely
- 15. 15 DSS - Cardiac Ailments • Tachycardia (Supraventricular and Ventricular) • Bradycardia • Premature Ventricular Contraction (PVC) • R on T • Various types of blocks (SA block, AV block, LBBB, RBBB) • ST segment changes (Ischemia, MI) • Fibrillation • Electrical Axis Deviations
- 16. 16 DSS for PVC • PVC stands for Premature Ventricular Contraction • This implies Ventricular Contraction takes place much before Atrial finishes its function (Premature) • The algorithm is based on single lead of information (Modified lead II information from MIT Arrhythmia data base) • Record Length 30 minutes • Samplingrate:360/samples/sec
- 17. 17 Features of PVCs • Change in RR interval • Absence of P wave • Bizarre Wave shape • Unifocal or Multifocal • Can occur in groups of one’s or two’s
- 18. 18 Occurrence of PVCs in groups of one’s PVC PVCNormal Normal Downwards T wave
- 19. 19 Occurrence of PVCs in groups of two’s PVC PVC PVC Two’sOne’s
- 20. 20 Detection Strategy • Arrhythmia Detection (Tompkin’s Algorithm) • Detection of Absence of P waves • Detection of Bizarre Wave shape using Hermite transforms • Similarity measurements using coherence function
- 22. 22 DSS Agents The functions of various DSS agents • Detection of QRS complex • Measurement of RR interval and detection of Arrhythmia • Detection of bizarre QRS complexes • M-shape detection • Negative peak detection • Detection of bi-phase QRS complexes • Detection of fibrillation related oscillations • Computation of Hermite transform of QRS complexes • ST segment measurements • Computation of KL Transform One or more agents facilitate detection of cardiac ailments
- 23. 23 Presentation Roadmap Introduction Use Case Scenarios Components Implementation and Demo
- 24. 24 Implementation • Stand-alone DSS (PVC) on MATLAB, PC and 64x • Remote Video Consultation System on PC and DM642 Patient @ Ambulance • Ambulance side system on 1. DSS on Laptop with Video Camera 2. DSS on DM642 with Video Camera • Doctor’s Viewer on PC
- 26. 26 Value-Add • 90% to 95% sensitivity and specificity on MIT-BIH database • Embedded DSS optimized to work with low resource requirements • Analysis by Embedded DSS in real-time and meant to pre warn catastrophic and premonitory arrhythmias
- 27. 27 Future Work Implementation • Porting of other cardiac disease detection functionalities of DSS on DSP • Porting of DSS on DA224 Ultra-low-power DSP Research • Application of wavelet transforms and Hermite functions in the detection of more cardiac ailments • Application of Self Organizing Maps in the classification of cardiac ailments • Agent based DSS
- 28. 28 Future Work - Agent Based DSS • A set of sensor nodes • Each element of the set executing the function of a specific agent • Will aid in the detection of cardiac ailment through collaborative processing • Coordination through high level Petri Nets Supporting dynamic changes
- 29. 29 Reference Database • MIT Data base – www.physionet.com • European ST Data base - www.physionet.com
- 30. 30 Thank you! Contact: arpan.pal@tcs.com Any questions, please?
Editor's Notes
- <number> Cardio vascular events affect 40 million people yearly and 1 in 3 deaths in the world is heart related. Time is of the essence in diagnosing and treating cardio vascular diseases. Delays can lead to increased morbidity and mortality and cause significant anxiety for patients. Real-time monitoring, analysis and response for mobile cardiac patients at home or away is the challenge today. The scenarios depicted below are the present challenges: Scenario 1 (Patient at Home) A patient at home with history of cardio vascular disease without any real-time ECG monitoring system. The patient suffers a cardiac attack and couldn’t even raise an alarm before succumbing to death. Scenario 2 (Patient at Ambulance) A patient who has suffered a massive cardiac arrest is being moved to the hospital in an ambulance which is located 30 minutes away. The doctor in the hospital can only attend the patient’s vital statistics such as ECG data, blood pressure etc., analyze the same and respond to the required treatment only after the patient reaches the hospital.
- <number> Now let us see how the previous TCS components are integrated to address the problems / scenarios defined in the first slide. Scenario 1 (Patient at Home) Patient’s Side The patient is wearing the 1/3/12 leads wearable ECG recorder. Optional DSS for alerting the patient and initiating data transmission with doctor Doctor’s side Provision to store and forward ECG data Real-time viewing of ECG signals from patient at the Doctor’s end (Click the ECG video during presentation to start)
- <number> Scenario 2 (Patient at Ambulance) Ambulance Side The patient is wearing the ECG recorder leads (12 lead) Remote Video System is installed with hardware box, camera and sound accessories ECG front-end interface with the hardware box DSS is embedded in the hardware box ECG display with DSS alert for aid to Ambulance attendant Wireless connectivity Hospital side Remote video system client is installed in doctor’s laptop / desktop Doctor can view patient’s video and ECG data remotely at his laptop/desktop Doctor can instruct the ambulance attended in case of emergency through two-way audio link (Click the ECG video during presentation to start)
- <number> To address the above problems TCS has come up with an integrated solution that provides real-time digital acquisition of the ECG waveform, rendering of waveform to cardiologist on-screen, remote data and patient monitoring, decision support system to aid the cardio vascular analysis in such scenarios and results in faster response from the cardiologists. The solution uses the following components built by TCS: 1/3/12 Lead wearable ECG Recorder with mobile connectivity Lightweight and portable Ultra low power Single Channel / Three channel option Mobile (GSM) connectivity 12 Lead ECG Recorder Single channel 12 lead ECG Acquisition Integrated thermal printer Zigbee enabled ECG Decision Support System Arrhythmia Detection Future Support for Tachycardia, Bradycardia, Asystole, Fusion Tested with MIT-AH Database 90% – 95% accuracy compared to 40% to 60% accuracy available. Advanced DSP techniques to achieve this accuracy Remote Video Consultation System One-way video with two-way audio Interface with Sensor instruments Highly compressed video and audio using state-of-the-art compression techniques Low-bandwidth / High-quality Proprietary error resilience algorithms for robust performance