1Copyright Š 2014 Tata Consultancy Services Limited
Dr. Arpan Pal
Principal Scientist and Head of Research
Innovation Lab, Kolkata
Tata Consultancy Services
Mobile Phone Sensing Based Affordable
Wellness/Healthcare
17 May 2015
2
Agenda
Status and Opportunity
Algorithms and Results
Mobile Sensing Health Solution
3
Click to edit Master title styleMobile Sensing Health Solution
4
Power in the Palm of People
ZuseZ3 1941
I think there will be a world market for
about five computers
- Thomas Watson, Chairman of IBM, 1941
The Reach
The Power
640KB of memory ought to be enough for
anybody -- - Bill Gates, in 1981
Windows Phone 2014
Quad Core Processor; 2048 MB
RAM; A Convergence Device
IBM PC in 1981
Smart phones to dwarf world population
The Value
FDA predicts more than 50% smart phones users will download a healthcare
app by 2018 (1.7 Billion) - Plans to regulate some healthcare apps , but also
will permit some without premarket reviews
5
Motivation
• Connected Health has the potential of disrupting healthcare through
24x7 monitoring of physiological parameters leading towards preventive
care and newer predictive diagnosis - however cost of ownership of medical
devices is an issue
• Digital India project: Government to ensure that every Indian has
smartphone by 2019 – Economic Times, Aug 25 2014 (Rs. 3000 )
• Smartphones have sensors and processing power in them to turn them
into cloud-connected physiological measuring devices – no extra cost of
ownership - Camera, Microphone, Accelerometer, Gyroscope, Magnetometer, GPS
Need to increase reach, inclusivity, affordability
Use Cases
Elderly People
Monitoring
@Home
Chronic Patient
Monitoring
@Home
Rural
Healthcare
Insurance
6
Mobile-Health - as it is now
Health Center/Home
ECG
Blood Pressure Monitor
Pulse OxyMeter
Mobile phone as
medical gateway
TCUP
Web Request
Patient
Records
Social
Network
Healthcare
Portal
Expert Doctor
7
Mobile-Health - proposed
Health Center/Home
Accelerometer
Microphone Camera …
on mobiles
TCUP
Web Request
Patient
Records
Social
Network
Healthcare
Portal
Expert Doctor
 Replace Medical Sensors with
Mobile Phone Sensing
 Multi-sensor Fusion
 Also explore wearables and
cheap mobile phone
attachments
8
Fall Detection
PPG extraction
Eye Image / Video
Cardiovascular
Model
Pulse Oxymetry
Pupilometry
Fingertip
Video
Lung Function
Blood Pressure
Microphone
Accelerometer
Digital Stethoscope
Heart Rate
Work @ TCS Research
Activity / Calorie
ECG
Respiratory Rate
HRV / Stress
Retinopathy
Pathology
Laryngeal Cancer
Camera Sensing
Uses TCUP, the
TCS IOT
Platform
9
Algorithms and Results
10
Heart Rate from PPG
 Sources of noise:
 improper finger placement
 imparting excessive pressure
 finger movement
11
Blood Pressure from PPG
2 Element Windkessel Model
12
Acoustic Digital Stethoscope and HRV using Mobile Phone Mic
• Stress Analysis from HRV - The sampling rate for audio is high compared to
PPG, thus calculation of HRV would be more accurate. From HRV, Stress Levels
can be estimated.
• Sound analysis - Analysis of normal and abnormal sounds or heart
• Fetal heart beat monitoring – The mobile phone mic is placed on lower abdomen.
Simultaneous capture of mother’s PPG can be used to adaptively filter mother’s
heart beat. Need to get access to hospitals for collecting data.
13
Mobile Phone based Activity Detection for Wellness
Activity Detection
– Uses Accelerometer Data
– Gyroscope and Magnetometer for orientation
correction
– Step Count, Stride Length Estimation
– Walking, Brisk Walking, Running Classification
Continuous Data Stream
Windowed Data
Zero Normalization
Linear Interpolation
Low Pass Filtration
Frequency Spectrum
Identifying non-activity window
using frequency spectrum
Peak Detection and Step
Validation using IPA;
calculating step cycle lengths for
all valid steps in the window
Classification of window
activity using step frequencies
derived from step cycle lengths
Continuous Data Stream
Windowed Data
Zero Normalization
Linear Interpolation
Low Pass Filtration
Frequency Spectrum
Identifying non-activity window
using frequency spectrum
Peak Detection and Step
Validation using IPA;
calculating step cycle lengths for
all valid steps in the window
Classification of window
activity using step frequencies
derived from step cycle lengths
Peak Detection and Step Validation
Calculating Step cycle lengths for all valid
steps in the window
Classification of window activity using step
frequencies derived from step cycle lengths
14
Camera – Heart Rate and Blood Pressure
Accelerometer - Activity Detection and Calorie Count
Microphone – Heart Sound
Some Results
Encouraging initial results for Pulse Oxymetry, HRV / Stress, Fall Detection
15
Experience certainty.
Status and Opportunity
16
Current Status
• Ongoing
Trials with TCS Doctors
• Activity Tracking, Heart Rate – Feb 2015
Trials within TCS Fit4Life
• Heart Rate, BP, HRV, Digital Stethoscope
Planned trials at Rural Chhattisgarh and
West Bengal
BP App won the best demo in Sensys 2014
at Memphis, USA
17
The opportunity…
• Mobile phone based physiological sensing for cloud based
remote healthcare can be the answer
India is said to have 1 doctor for 1700 people.
But is every doctor reaching 1700 people?
• Monitoring in rural PHCs, in partnership with the government
• Regular monitoring based preventive healthcare for insurance
• Elderly people monitoring for old age homes
• Bundled offerings with Telecom providers
• Inclusive and affordable Telemedicine for hospitals
The Business?
• Collaborate with local doctors and rural health centers
• Address India-specific issues and problems
How to go about it?
18
1. "Mobile healthcare infrastructure for home and small clinic" ACM international workshop on
Pervasive Wireless Healthcare, 2012.
2. "A robust heart rate detection using smart-phone video" ACM MobiHoc workshop on
Pervasive wireless healthcare, 2013.
3. "UbiHeld: ubiquitous healthcare monitoring system for elderly and chronic patients”, ACM
conference on Pervasive and ubiquitous computing, 2013.
4. "HeartSense: estimating blood pressure and ECG from photoplethysmograph using smart
phones”, ACM Conference on Embedded Networked Sensor Systems, 2013.
5. "Estimation of blood pressure levels from reflective Photoplethysmograph using smart
phones." ACM BIBE 2013.
6. "Estimation of ECG parameters using photoplethysmography." ACM BIBE 2013
7. "PhotoECG: Photoplethysmography to Estimate ECG Parameters" IEEE ICASP, 2014
8. "Improved Heart Rate Detection using Smart-phone" ACM SAC, 2014.
9. "Estimating Blood Pressure using Windkessel Model on Photoplethysmogram" EMBC 2014.
10. "Smart Phone Based Blood Pressure Indicator" MobileHealth workshop of Mobihoc 2014
11. "HeartSense: Estimating Heart rate from Smartphone Photoplethysmogram using Adaptive
Filter and Interpolation" HealthyIoT, IoT-360, 2014
12. "AcTrak - Unobtrusive Activity Detection and Step Counting using Smartphones" Mobiquitous
2013
Published Papers
19
Thank You
arpan.pal@tcs.com

Icdcn industry track_arpanpal

  • 1.
    1Copyright Š 2014Tata Consultancy Services Limited Dr. Arpan Pal Principal Scientist and Head of Research Innovation Lab, Kolkata Tata Consultancy Services Mobile Phone Sensing Based Affordable Wellness/Healthcare 17 May 2015
  • 2.
    2 Agenda Status and Opportunity Algorithmsand Results Mobile Sensing Health Solution
  • 3.
    3 Click to editMaster title styleMobile Sensing Health Solution
  • 4.
    4 Power in thePalm of People ZuseZ3 1941 I think there will be a world market for about five computers - Thomas Watson, Chairman of IBM, 1941 The Reach The Power 640KB of memory ought to be enough for anybody -- - Bill Gates, in 1981 Windows Phone 2014 Quad Core Processor; 2048 MB RAM; A Convergence Device IBM PC in 1981 Smart phones to dwarf world population The Value FDA predicts more than 50% smart phones users will download a healthcare app by 2018 (1.7 Billion) - Plans to regulate some healthcare apps , but also will permit some without premarket reviews
  • 5.
    5 Motivation • Connected Healthhas the potential of disrupting healthcare through 24x7 monitoring of physiological parameters leading towards preventive care and newer predictive diagnosis - however cost of ownership of medical devices is an issue • Digital India project: Government to ensure that every Indian has smartphone by 2019 – Economic Times, Aug 25 2014 (Rs. 3000 ) • Smartphones have sensors and processing power in them to turn them into cloud-connected physiological measuring devices – no extra cost of ownership - Camera, Microphone, Accelerometer, Gyroscope, Magnetometer, GPS Need to increase reach, inclusivity, affordability Use Cases Elderly People Monitoring @Home Chronic Patient Monitoring @Home Rural Healthcare Insurance
  • 6.
    6 Mobile-Health - asit is now Health Center/Home ECG Blood Pressure Monitor Pulse OxyMeter Mobile phone as medical gateway TCUP Web Request Patient Records Social Network Healthcare Portal Expert Doctor
  • 7.
    7 Mobile-Health - proposed HealthCenter/Home Accelerometer Microphone Camera … on mobiles TCUP Web Request Patient Records Social Network Healthcare Portal Expert Doctor  Replace Medical Sensors with Mobile Phone Sensing  Multi-sensor Fusion  Also explore wearables and cheap mobile phone attachments
  • 8.
    8 Fall Detection PPG extraction EyeImage / Video Cardiovascular Model Pulse Oxymetry Pupilometry Fingertip Video Lung Function Blood Pressure Microphone Accelerometer Digital Stethoscope Heart Rate Work @ TCS Research Activity / Calorie ECG Respiratory Rate HRV / Stress Retinopathy Pathology Laryngeal Cancer Camera Sensing Uses TCUP, the TCS IOT Platform
  • 9.
  • 10.
    10 Heart Rate fromPPG  Sources of noise:  improper finger placement  imparting excessive pressure  finger movement
  • 11.
    11 Blood Pressure fromPPG 2 Element Windkessel Model
  • 12.
    12 Acoustic Digital Stethoscopeand HRV using Mobile Phone Mic • Stress Analysis from HRV - The sampling rate for audio is high compared to PPG, thus calculation of HRV would be more accurate. From HRV, Stress Levels can be estimated. • Sound analysis - Analysis of normal and abnormal sounds or heart • Fetal heart beat monitoring – The mobile phone mic is placed on lower abdomen. Simultaneous capture of mother’s PPG can be used to adaptively filter mother’s heart beat. Need to get access to hospitals for collecting data.
  • 13.
    13 Mobile Phone basedActivity Detection for Wellness Activity Detection – Uses Accelerometer Data – Gyroscope and Magnetometer for orientation correction – Step Count, Stride Length Estimation – Walking, Brisk Walking, Running Classification Continuous Data Stream Windowed Data Zero Normalization Linear Interpolation Low Pass Filtration Frequency Spectrum Identifying non-activity window using frequency spectrum Peak Detection and Step Validation using IPA; calculating step cycle lengths for all valid steps in the window Classification of window activity using step frequencies derived from step cycle lengths Continuous Data Stream Windowed Data Zero Normalization Linear Interpolation Low Pass Filtration Frequency Spectrum Identifying non-activity window using frequency spectrum Peak Detection and Step Validation using IPA; calculating step cycle lengths for all valid steps in the window Classification of window activity using step frequencies derived from step cycle lengths Peak Detection and Step Validation Calculating Step cycle lengths for all valid steps in the window Classification of window activity using step frequencies derived from step cycle lengths
  • 14.
    14 Camera – HeartRate and Blood Pressure Accelerometer - Activity Detection and Calorie Count Microphone – Heart Sound Some Results Encouraging initial results for Pulse Oxymetry, HRV / Stress, Fall Detection
  • 15.
  • 16.
    16 Current Status • Ongoing Trialswith TCS Doctors • Activity Tracking, Heart Rate – Feb 2015 Trials within TCS Fit4Life • Heart Rate, BP, HRV, Digital Stethoscope Planned trials at Rural Chhattisgarh and West Bengal BP App won the best demo in Sensys 2014 at Memphis, USA
  • 17.
    17 The opportunity… • Mobilephone based physiological sensing for cloud based remote healthcare can be the answer India is said to have 1 doctor for 1700 people. But is every doctor reaching 1700 people? • Monitoring in rural PHCs, in partnership with the government • Regular monitoring based preventive healthcare for insurance • Elderly people monitoring for old age homes • Bundled offerings with Telecom providers • Inclusive and affordable Telemedicine for hospitals The Business? • Collaborate with local doctors and rural health centers • Address India-specific issues and problems How to go about it?
  • 18.
    18 1. "Mobile healthcareinfrastructure for home and small clinic" ACM international workshop on Pervasive Wireless Healthcare, 2012. 2. "A robust heart rate detection using smart-phone video" ACM MobiHoc workshop on Pervasive wireless healthcare, 2013. 3. "UbiHeld: ubiquitous healthcare monitoring system for elderly and chronic patients”, ACM conference on Pervasive and ubiquitous computing, 2013. 4. "HeartSense: estimating blood pressure and ECG from photoplethysmograph using smart phones”, ACM Conference on Embedded Networked Sensor Systems, 2013. 5. "Estimation of blood pressure levels from reflective Photoplethysmograph using smart phones." ACM BIBE 2013. 6. "Estimation of ECG parameters using photoplethysmography." ACM BIBE 2013 7. "PhotoECG: Photoplethysmography to Estimate ECG Parameters" IEEE ICASP, 2014 8. "Improved Heart Rate Detection using Smart-phone" ACM SAC, 2014. 9. "Estimating Blood Pressure using Windkessel Model on Photoplethysmogram" EMBC 2014. 10. "Smart Phone Based Blood Pressure Indicator" MobileHealth workshop of Mobihoc 2014 11. "HeartSense: Estimating Heart rate from Smartphone Photoplethysmogram using Adaptive Filter and Interpolation" HealthyIoT, IoT-360, 2014 12. "AcTrak - Unobtrusive Activity Detection and Step Counting using Smartphones" Mobiquitous 2013 Published Papers
  • 19.