Robust Lowcost NonIntrusive Sensing and Recognition of Seated
Robust, Low-cost, Non-Intrusive Sensing and Recognition of Seated Postures Bilge Mutlu, Andreas Krause, Jodi Forlizzi, Carlos Guestrin, and Jessica Hodgins Human-Computer Interaction Institute, Carnegie Mellon University
Why seated postures? Automobile Home Classroom Wheelchair Office
Using posture information Today’s talk
Existing approaches • Kinesthetic Motion-capture markers or conductiveelastomer-embedded fabrics Pellegrini and Iocchi. , 2006
Existing approaches • Kinesthetic Motion-capture markers or conductiveelastomer-embedded fabrics • Vision-based Image sequences from a single camera or multiple cameras Tognetti et al. , 2005
Existing approaches • Kinesthetic Motion-capture markers or conductiveelastomer-embedded fabrics • Vision-based Image sequences from a single camera or multiple cameras • Pressure-sensing-based Pressure readings from the seating surfaces Han et al. , 2001
Challenges Robust generalization • Poor generalization Good performance in classifying “familiar” subjects, poor performance with “unfamiliar” subjects due to high dimensionality. • High cost Low-cost High-fidelity pressure sensors are expensive. • Slow performance Near-real-time performance Processing high-fidelity sensor data demands computational power, which leads to slow processing.
Our solution • Robust generalization Up to 87% accuracy in classifying 10 postures with new subjects. • Low cost Using 19 pressure sensors instead of 4032. Reducing sensor cost from $3 K to ~$100. • Near-real-time performance 10 Hz on a standard desktop computer • Novel methodology Using domain knowledge and nearoptimal sensor placement.
Methodology
Learning Algorithm • Logistic Regression Sparse representation ✴ • Cross-validation 10 -fold, gender-balanced training and testing samples from different subjects • Separate sets Training, 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.
Modeling Understanding pressure data
Modeling Understanding pressure data
Modeling Understanding our data
Modeling Domain knowledge
Modeling Features Size and position of bounding boxes Distances to the edges of the seat Distance and angle to between bounding boxes Pressure applied to the bottom area Parameters of the ellipses that fit the bottom area
Modeling Features Classification accuracy
Modeling Separability test
Modeling Feature elimination
Methodology
Dimensionality Reduction Sensor granularity
Dimensionality Reduction Sensor granularity
Dimensionality Reduction 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 F V A⊆V IG(A; F) = H(F) - H(F | A)
Dimensionality Reduction Near-optimal placement
Dimensionality Reduction Sensor placements
Dimensionality Reduction Near-optimal placement Classification accuracy
Methodology
Prototyping
Evaluation of prototype • 20 naive participants 10 -fold cross validation testing with %5 of the data • 78% accuracy In classifying 10 postures • 10 Hz real-time performance On a standard desktop computer
Methodology
Conclusions • Generalizability Up to 87% (with a base rate of 10%) achieved with unfamiliar subjects. • Low cost Higher classification accuracy than existing systems using less than 1% of the sensors. ~ $100 sensor cost compared to the commercial sensor for $3 K (33 times reduction in price). • Near-real-time performance At 10 Hz on a standard desktop computer.
Applications Automobile Home Classroom Wheelchair Office
Next Steps Future challenges • Transferring learning across chairs A “transformation map” could be created • Only static postures Temporal dimension needs to be considered • The set of ten postures The set of postures should come from the activity
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 Next Steps 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
- Slides: 34