Mitja Luštrek Božidara Cvetković and Simon Kozina Jožef Stefan Institute Department of Intelligent Systems Slovenia Introduction Motivation Chiron project monitoring of congestive heart failure patients ID: 775820
Download Presentation The PPT/PDF document " Energy expenditure estimation with wea..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
Energy expenditure estimation with wearable accelerometers
Mitja Luštrek,
Božidara Cvetković and Simon Kozina
Jožef Stefan Institute
Department of Intelligent Systems
Slovenia
Slide2Introduction
Motivation:Chiron project – monitoring of congestive heart failure patientsThe patient’s energy expenditure (= intensity of movement) provides context for heart activity
Slide3Introduction
Motivation:Chiron project – monitoring of congestive heart failure patientsThe patient’s energy expenditure (= intensity of movement) provides context for heart activity Method:Two wearable accelerometers → accelerationAcceleration → activityAcceleration + activity → energy expenditure
Machine
learning
Slide4Measuring human energy expenditure
Direct calorimetry
Heat output of the patient
Most reliable
,
laboratory conditions
Slide5Measuring human energy expenditure
Direct calorimetry
Heat output of the patient
Most reliable
,
laboratory conditions
Indirect calorimetry
Inhaled and exhaled oxygen and CO
2
Quite reliable
,
field conditions
,
mask needed
Slide6Measuring human energy expenditure
Direct calorimetry
Heat output of the patient
Most reliable
,
laboratory conditions
Indirect calorimetry
Inhaled and exhaled oxygen and CO
2
Quite reliable
,
field conditions
,
mask needed
Diary
Simple
,
Unreliable
,
patient-dependant
Slide7Measuring human energy expenditure
Direct calorimetry Heat output of the patient Most reliable, laboratory conditionsIndirect calorimetry Inhaled and exhaled oxygen and CO2 Quite reliable, field conditions, mask neededDiary Simple, Unreliable, patient-dependant
Wearable accelerometers
Slide8Hardware
Co-located with ECG
One placement
to be selected
Slide9Hardware
Co-located with ECG
One placement
to be selected
Shimmer sensor nodes
3-axial accelerometer @ 50 Hz
Bluetooth and 802.15.4 radio
Microcontroller
Custom firmware
Slide10Hardware
Co-located with ECG
One placement
to be selected
Shimmer
sensor nodes
3-axial accelerometer @ 50 Hz
Bluetooth and 802.15.4 radio
Microcontroller
Custom firmware
Android smartphone
Bluetooth
Slide11Training/test data
Activity
Lying
Sitting
Standing
Walking
Running
Cycling
Scrubbing the floor
Sweeping
...
Slide12Training/test data
ActivityEnergy expenditureLying1.0 METSitting1.0 METStanding1.2 METWalking3.3 METRunning11.0 METCycling8.0 METScrubbing the floor3.0 METSweeping4.0 MET...
1 MET = energyexpendedat rest
Recorded
by five
volunteers
Slide13Machine learning procedure
atat+1at+2...
Acceleration data
Sliding window (2 s)
Slide14Machine learning procedure
atat+1at+2...
Acceleration data
Sliding window (2 s)
f
1
f
2
f
3
...
Activity
Training
Machine learning
AR Classifier
Slide15Machine learning procedure
atat+1at+2...
Acceleration data
f
1f2f3...
Use/testing
Activity
Sliding window (2 s)
AR Classifier
Slide16Machine learning procedure
atat+1at+2...
Acceleration data
Activity
AR Classifier
Slide17Machine learning procedure
atat+1at+2...
Acceleration data
Sliding window (10 s)
Activity
AR Classifier
Slide18Machine learning procedure
atat+1at+2...
Acceleration data
Sliding window (10 s)
f’
1
f’
2
f’
3
...
Activity
EE
Training
Machine learning (regression)
EEE Classifier
Activity
AR Classifier
Slide19Machine learning procedure
atat+1at+2...
Acceleration data
Sliding window (10 s)
f’
1
f’
2
f’
3
...
Activity
Use/testing
EEE Classifier
Activity
AR Classifier
EE
Slide20Machine learning procedure
atat+1at+2...
Acceleration data
EE
Energy expenditure
Slide21Features for activity recognition
Average acceleration
Variance in acceleration
Minimum and maximum acceleration
Speed of change between min. and max.
Accelerometer orientation
Frequency domain features (FFT)
Correlations between accelerometer axes
Slide22Features for energy expenditure est.
ActivityAverage length of the acceleration vectorNumber of peaks and bottoms of the signal
Slide23Features for energy expenditure est.
ActivityAverage length of the acceleration vectorNumber of peaks and bottoms of the signalArea under accelerationArea under gravity-subtracted acceleration
Slide24Features for energy expenditure est.
Activity
Average length of the acceleration vector
Number of peaks and bottoms of the signal
Area under acceleration
Area under
gravity-subtracted
acceleration
Change in velocity
Change in kinetic energy
Slide25Sensor placement and algorithm
Linear regressionSupport vector regressionRegression treeModel treeNeural networkChest + ankle5.093.291.412.181.65Chest + thigh6.753.681.582.381.66Chest + wrist6.753.941.304.951.39
Mean absolute error in MET
Slide26Sensor placement and algorithm
Linear regressionSupport vector regressionRegression treeModel treeNeural networkChest + ankle5.093.291.412.181.65Chest + thigh6.753.681.582.381.66Chest + wrist6.753.941.304.951.39
Mean absolute error in MET
Slide27Sensor placement and algorithm
Linear regressionSupport vector regressionRegression treeModel treeNeural networkChest + ankle5.093.291.412.181.65Chest + thigh6.753.681.582.381.66Chest + wrist6.753.941.304.951.39
Mean absolute error in MET
Slide28Sensor placement and algorithm
Linear regressionSupport vector regressionRegression treeModel treeNeural networkChest + ankle5.093.291.412.181.65Chest + thigh6.753.681.582.381.66Chest + wrist6.753.941.304.951.39
Mean absolute error in MET
Slide29Sensor placement and algorithm
Linear regressionSupport vector regressionRegression treeModel treeNeural networkChest + ankle5.093.291.412.181.65Chest + thigh6.753.681.582.381.66Chest + wrist6.753.941.304.951.39
Mean absolute error in MET
Slide30Sensor placement and algorithm
Linear regressionSupport vector regressionRegression treeModel treeNeural networkChest + ankle5.093.291.412.181.65Chest + thigh6.753.681.582.381.66Chest + wrist6.753.941.304.951.39
Mean absolute error in MET
Slide31Sensor placement and algorithm
Linear regressionSupport vector regressionRegression treeModel treeNeural networkChest + ankle5.093.291.412.181.65Chest + thigh6.753.681.582.381.66Chest + wrist6.753.941.304.951.39
Mean absolute error in MET
Lowest error, poor extrapolation, interpolation
Second lowest error, better flexibility
Slide32Estimated vs. true energy
Average
error:
1.39 MET
Slide33Estimated vs. true energy
Low intensity
Moderate
intensity
Running, cycling
Average
error:
1.39 MET
Slide34Estimated vs. true energy
Low intensity
Moderate
intensity
Running, cycling
Average
error:
1.39 MET
Slide35Multiple classifiers
Activity
AR Classifier
Slide36Multiple classifiers
Activity
AR Classifier
General
EEE Classifier
EE
Cycling
EEE Classifier
RunningEEE Classifier
Activity = cycling
Activity = running
Activity = other
Slide37Estimated vs. true energy, multiple cl.
Low intensity
Moderate
intensity
Running, cycling
Average
error:
0.91 MET
Slide38Conclusion
Energy expenditure estimation with wearable accelerometers using machine learning
Study of sensor placements and algorithms
Multiple classifiers: error 1.39 → 0.91 MET
Slide39Conclusion
Energy expenditure estimation with wearable accelerometers using machine learning
Study of sensor placements and algorithms
Multiple classifiers: error 1.39 → 0.91 MET
Cardiologists judged suitable to monitor congestive heart
failure
patients
Other medical and sports applications possible