7
WLSA CONVERGENCE SUMMIT MOTION ASSESSMENT FOR ROBOTIC SURGERY EDUCATION USING INERTIAL BODY SENSORS JIAQI GONG, JOHN LACH CHARLES L. BROWN DEPARTMENT OF ELECTRICAL AND COMPUTER ENGINEERING UNIVERSITY OF VIRGINIA

WH2014 Session: Motion assessment for robotic surgery education using inertial body sensors

Embed Size (px)

DESCRIPTION

From Wireless Health 2014 Demo and Abstract Presentation Session 1, featuring speaker Jiaqi Gong.

Citation preview

Page 1: WH2014 Session: Motion assessment for robotic surgery education using inertial body sensors

WLSACONVERGENCE SUMMIT

MOTION ASSESSMENT FOR ROBOTIC SURGERY EDUCATION USING INERTIAL BODY SENSORS

JIAQI GONG, JOHN LACHCHARLES L. BROWN DEPARTMENT OF ELECTRICAL AND COMPUTER ENGINEERINGUNIVERSITY OF VIRGINIA

Page 2: WH2014 Session: Motion assessment for robotic surgery education using inertial body sensors

MOTION ASSESSMENT FOR ROBOTIC SURGERY EDUCATION USING

INERTIAL BODY SENSORS

Wireless Health, October 30th, 2014

Jiaqi Gong, John Lach

Charles L. Brown Department of Electrical and

Computer EngineeringUniversity of Virginia

Yanjun QiDepartment of

Computer ScienceUniversity of Virginia

Rebecca S. Zee, Sierra J. Seaman,

Noah S. Schenkman

Department of UrologyUniversity of Virginia

UVA CENTER FORWIRELESS HEALTH

Page 3: WH2014 Session: Motion assessment for robotic surgery education using inertial body sensors

Background and Purpose

0 0.5 1 1.5 2 2.5 3 3.5 4

x 104

-1000

0

1000Right Wrist

0 0.5 1 1.5 2 2.5 3 3.5 4

x 104

-1000

0

1000Left Wrist

0 0.5 1 1.5 2 2.5 3 3.5 4

x 104

-1000

0

1000Right Upper Arm

0 0.5 1 1.5 2 2.5 3 3.5 4

x 104

-1000

0

1000Left Ankle

Score

Data

Feedback

Page 4: WH2014 Session: Motion assessment for robotic surgery education using inertial body sensors

Methods: (Modeling and Clustering)

S1

S2

Sk

… … … … …

Raw Data܇ 1 ൌ��ሺ1 ǡ��1 ǥ (1

ǥ

1ሺ1ሻא ǥ

1ሺ1ሻא ǥ

אሺ1ሻ1 t Samples

PreprocessingԢ ሺ1ሻൌ��ሺ1Ԣ ǡ��1Ԣ ǥ 1Ԣ)

t Samples

ǥ

1Ԣሺ1ሻא ǥ

1Ԣ ሺ1ሻא ǥ

1Ԣሺ1ሻא

SegmentationWൌ��ሼ1 ǡ��1 ǥ 1 ሽ

ǥ

ǥ

ǥ

1 1 ǥ

m Segments

Feature Extractionאሾ ǡ�� ሿൌ��ሼ1 ǡ��1 ǡ��ǥ 1ሽ11

11

1 1

Clusteringሺ1ǡ�ǡ��ሻሺ1 ǡ� ሻሺ1 ǡ� ሻ

ሺ1ǡ� ሻ

11

Class

m Segments C Classes

1. Segmentation based on 3D Curvature-Gradient Invariant to human factors during the mounting process2. Feature Extraction based on Control System Theory Invariant to observation variances during the sensing process3. Intra-subject Clustering Discover the motion pieces in each time-series data4. Inter-subject Clustering based on Dictionary Method and Prior Knowledge Discover the motion patterns which can be used to differentiate experts and novices

Page 5: WH2014 Session: Motion assessment for robotic surgery education using inertial body sensors

Results

Body-worn inertial sensors capturing motion and posture data provide better surgical skills assessments and more specific feedback than simulator performance alone. P Value Robotic

SimulatorInertial BSN

EIF P<0.5 P<0.05

EOM P<0.05 P<0.05

NSMP N/A P<0.0001

Overall P<0.05 N/A

Page 6: WH2014 Session: Motion assessment for robotic surgery education using inertial body sensors

Contact Us

Contact us with any question you have Jiaqi Gong ([email protected])

UVA CENTER FORWIRELESS HEALTH

Page 7: WH2014 Session: Motion assessment for robotic surgery education using inertial body sensors

WLSACONVERGENCE SUMMIT

www.wirelesshealth2014.org