Bilge Mutlu, Andreas Krause, Jodi Forlizzi, Carlos Guestrin, and Jessica Hodgins Human-Computer Interaction Institute, Carnegie Mellon University Robust, Low-cost, Non-Intrusive Sensing and Recognition of Seated Postures
Dec 30, 2015
Bilge Mutlu, Andreas Krause, Jodi Forlizzi, Carlos Guestrin, and Jessica Hodgins
Human-Computer Interaction Institute, Carnegie Mellon University
Robust, Low-cost, Non-Intrusive Sensing and Recognition of Seated
Postures
Why seated postures?
AutomobileAutomobile ClassroomClassroom WheelchairWheelchair
HomeHome OfficeOffice
Using posture information
Today’s talk
Pellegrini and Iocchi., 2006
• KinestheticMotion-capture markers or conductive- elastomer-embedded fabrics
Existing approaches
• KinestheticMotion-capture markers or conductive- elastomer-embedded fabrics
• Vision-basedImage sequences from a single camera or multiple cameras
Tognetti et al., 2005
Existing approaches
• KinestheticMotion-capture markers or conductive- elastomer-embedded fabrics
• Vision-basedImage sequences from a single camera or multiple cameras
• Pressure-sensing-basedPressure readings from the seating surfaces
Existing approaches
Han et al., 2001
•Poor generalizationGood performance in classifying “familiar” subjects, poor performance with “unfamiliar” subjects due to high dimensionality.
•High costHigh-fidelity pressure sensors are expensive.
•Slow performanceProcessing high-fidelity sensor data demands computational power, which leads to slow processing.
ChallengesRobust generalization
Low-cost
Near-real-time performance
Our solution•Robust generalization
Up to 87% accuracy in classifying 10 postures with new subjects.
•Low costUsing 19 pressure sensors instead of 4032. Reducing sensor cost from $3K to ~$100.
•Near-real-time performance10Hz on a standard desktop computer
•Novel methodologyUsing domain knowledge and near-optimal sensor placement.
Methodology
Learning Algorithm•Logistic Regression
Sparse representation
•Cross-validation10-fold, gender-balanced training and testing samples from different subjects
•Separate setsTraining, testing, and reporting samples from 52 people in 5 trials
• Implementation in Java
We would like to thank Hong Tan and Lynne Slivovsky for providing their data set for comparison.
✴
Understanding pressure dataModeling
Understanding pressure dataModeling
Understanding our dataModeling
Domain knowledgeModeling
FeaturesModeling
Size and position of bounding
boxes
Distances to the edges of the
seat
Distance and angle to between bounding
boxes
Parameters of the ellipses that fit the bottom
area
Pressure applied to the bottom
area
FeaturesModeling
Classification accuracy
Separability testModeling
Feature eliminationModeling
Methodology
Dimensionality Reduction
Sensor granularity
Dimensionality Reduction
Sensor granularity
How to place sensors?•F, feature variables
•V, locations and granularities
•A subset A of V that maximizes information gain about F where H is entropy
•NP-Hard optimization problem
•We use near-optimal approximation algorithm
Dimensionality Reduction
IG(A;F) = H(F) - H(F | A)
FV
A ⊆ V
Near-optimal placementDimensionality Reduction
Sensor placementsDimensionality Reduction
Near-optimal placementDimensionality Reduction
Classification accuracy
Methodology
Prototyping
Evaluation of prototype• 20 naive participants
10-fold cross validation testing with %5 of the data
• 78% accuracyIn classifying 10 postures
• 10 Hz real-time performanceOn a standard desktop computer
Methodology
Conclusions•Generalizability
Up to 87% (with a base rate of 10%) achieved with unfamiliar subjects.
•Low costHigher classification accuracy than existing systems using less than 1% of the sensors. ~ $100 sensor cost compared to the commercial sensor for $3K (33 times reduction in price).
•Near-real-time performanceAt 10Hz on a standard desktop computer.
Applications
AutomobileAutomobile ClassroomClassroom WheelchairWheelchair
HomeHome OfficeOffice
Future challenges•Transferring learning across chairs
A “transformation map” could be created
•Only static posturesTemporal dimension needs to be considered
•The set of ten posturesThe set of postures should come from the activity
Next Steps
Summary of Contributions• A non-intrusive, robust, low-cost system that recognizes
seated postures with generalizable, near-real-time performance.
• A novel methodology that uses domain-knowledge and near-optimal sensor placement strategy for classification.
This work was supported by NSF grants IIS-0121426, DGE- 0333420, CNS-0509383, Intel Corporation and Ford Motor Company.
From Postures to Activities• Reading the paper
• Watching TV
• Reading paperwork
• Watching TV + eating
• Sleeping
• Talking on the phone
• Reading a book
• Craftwork
• Reading the paper + watching TV
• Reading the paper + eating
Next Steps