Hips do lie! A position-aware mobile fall detection system Christian - - PowerPoint PPT Presentation

▶
hips do lie a position aware mobile fall detection system
SMART_READER_LITE
LIVE PREVIEW

Hips do lie! A position-aware mobile fall detection system Christian - - PowerPoint PPT Presentation

Apr 01, 2024 243 likes •614 views

Hips do lie! A position-aware mobile fall detection system Christian Krupitzer, Timo Sztyler, Janick Edinger, Martin Breitbach, Heiner Stuckenschmidt, and Christian Becker Proposition So be wise and keep on Reading the signs of my body And I'm

slide-1
SLIDE 1

Hips do lie! A position-aware mobile fall detection system

Christian Krupitzer, Timo Sztyler, Janick Edinger, Martin Breitbach, Heiner Stuckenschmidt, and Christian Becker

slide-2
SLIDE 2

Proposition

So be wise and keep on Reading the signs of my body And I'm on tonight you know my hips don't lie.

Hips Do Lie! A Position-Aware Mobile Fall Detection System

  • C. Krupitzer, T. Sztyler, J. Edinger, M. Breitbach,
  • H. Stuckenschmidt, and C. Becker

SHAKIRA: “HIPS DON’T LIE” (2005)

slide-3
SLIDE 3

3

Alternative Proposition

Hips Do Lie! A Position-Aware Mobile Fall Detection System

  • C. Krupitzer, T. Sztyler, J. Edinger, M. Breitbach,
  • H. Stuckenschmidt, and C. Becker

Observation: Hips might not be the most reliable source of information

slide-4
SLIDE 4

Research Question

How can a mobile fall detection system react to changes in the sensor position?

Hips Do Lie! A Position-Aware Mobile Fall Detection System

  • C. Krupitzer, T. Sztyler, J. Edinger, M. Breitbach,
  • H. Stuckenschmidt, and C. Becker

?

slide-5
SLIDE 5

5

Hips Do Lie! A Position-Aware Mobile Fall Detection System

  • C. Krupitzer, T. Sztyler, J. Edinger, M. Breitbach,
  • H. Stuckenschmidt, and C. Becker

Fall Detection in a Nutshell Self-Adaptive Fall Detection Implemen- tation Case Studies Future Work

What to Expect from this Presentation?

slide-6
SLIDE 6

Fall Detection Systems

Wearables

6

Hips Do Lie! A Position-Aware Mobile Fall Detection System

  • C. Krupitzer, T. Sztyler, J. Edinger, M. Breitbach,
  • H. Stuckenschmidt, and C. Becker

Vision-based Ambient-based

Mubashir, Shao & Seed: A survey on fall detection: Principles and approaches. In: Neurocomputing, Volume 100, 2013, Pages 144-152.

slide-7
SLIDE 7

7

Positions of Wearables for Fall Detection

Head Chest Waist Wrist Hip Ankle

Hips Do Lie! A position-aware mobile fall detection system

  • C. Krupitzer, T. Sztyler, J. Edinger, M. Breitbach,
  • H. Stuckenschmidt and C. Becker

Jump Sit Walk Bend Run Sit Walk Get up Jump Walk Fall

slide-8
SLIDE 8

Acc. Data

Caregivers Notification Adaptation Logic

Monitor Planner Executor Analyzer

8

Self-Adaptive Fall Detection

Data acquisition Data preprocessing Fall alert Fall detection

FD FA

Patient Sensors Acc.

Z Y X

Gyro Magn.

Hips Do Lie! A Position-Aware Mobile Fall Detection System

  • C. Krupitzer, T. Sztyler, J. Edinger, M. Breitbach,
  • H. Stuckenschmidt, and C. Becker

FD FA DP DP DA DA

slide-9
SLIDE 9

Acc. Data

Monitor Planner Executor Analyzer

9

Self-Adaptive Fall Detection

Data acquisition Data preprocessing Fall alert Fall detection

FD FA

Patient Sensors Acc.

Z Y X

Hips Do Lie! A Position-Aware Mobile Fall Detection System

  • C. Krupitzer, T. Sztyler, J. Edinger, M. Breitbach,
  • H. Stuckenschmidt, and C. Becker

FD FA DP DP DA DA FA FD DP DA

Adaptation Logic Caregivers Notification

slide-10
SLIDE 10

10

Implementation: Adaptation Logic

Hips Do Lie! A Position-Aware Mobile Fall Detection System

  • C. Krupitzer, T. Sztyler, J. Edinger, M. Breitbach,
  • H. Stuckenschmidt, and C. Becker

Planner Executor

  • Acc. Data:

(x,y,z) Trigger Fall Notification Fall Handling Notifi- cation

Monitor Analyzer

?

Fall Alert

  • Acc. Data (x,y,z)
slide-11
SLIDE 11

Window Manager

11

Implementation: Adaptation Logic

Hips Do Lie! A Position-Aware Mobile Fall Detection System

  • C. Krupitzer, T. Sztyler, J. Edinger, M. Breitbach,
  • H. Stuckenschmidt, and C. Becker

Planner Executor

  • Acc. Data:

(x,y,z) Trigger Fall Notification Fall Handling

Monitor Analyzer

Notifi- cation

t1: (x1,y1,z1) ... tn: (xn,yn,zn) ... tn+x: (xn+x,yn+x,zn+x)

w1 Feature1 Feature2 ... w2 Feature1 Feature2 ...

...

... ... ...

slide-12
SLIDE 12

Window Manager

12

Implementation: Adaptation Logic

Hips Do Lie! A Position-Aware Mobile Fall Detection System

  • C. Krupitzer, T. Sztyler, J. Edinger, M. Breitbach,
  • H. Stuckenschmidt, and C. Becker

Planner Executor

  • Acc. Data:

(x,y,z) Trigger Fall Notification Fall Handling

Monitor Analyzer

Event-based approach

Notifi- cation

Continuous approach

versus

slide-13
SLIDE 13

Window Manager

13

Implementation: Adaptation Logic

Hips Do Lie! A Position-Aware Mobile Fall Detection System

  • C. Krupitzer, T. Sztyler, J. Edinger, M. Breitbach,
  • H. Stuckenschmidt, and C. Becker

Planner Executor

  • Acc. Data:

(x,y,z) Trigger Fall Notification Fall Handling

Monitor Analyzer

Event-based approach

Notifi- cation

Continuous approach

versus

slide-14
SLIDE 14

Acc. Data

Caregivers Notification Adaptation Logic

Monitor Planner Executor Analyzer

14

Self-Adaptive Fall Detection

Data acquisition Data preprocessing Fall alert Fall detection

FD FA

Patient Sensors Acc.

Z Y X

Hips Do Lie! A Position-Aware Mobile Fall Detection System

  • C. Krupitzer, T. Sztyler, J. Edinger, M. Breitbach,
  • H. Stuckenschmidt, and C. Becker

DP DA DA FA FD DP

slide-15
SLIDE 15

15

Self-Improving Fall Detection

Hips Do Lie! A Position-Aware Mobile Fall Detection System

  • C. Krupitzer, T. Sztyler, J. Edinger, M. Breitbach,
  • H. Stuckenschmidt, and C. Becker

Monitor Planner Executor Analyzer

Adaptation Logic

Subject Analyzer Algorithm Chooser Algorithms A1 A2 An … Data Monitor Adaptation Executor Position Analyzer

Self-Improvement Module Data Instructions

Subject Clusterer

X-Means Age Body Mass Index (BMI) Sztyler 2016

Algorithm Optimizer

slide-16
SLIDE 16

16

Evaluation Setting

Hips Do Lie! A Position-Aware Mobile Fall Detection System

  • C. Krupitzer, T. Sztyler, J. Edinger, M. Breitbach,
  • H. Stuckenschmidt, and C. Becker

Training of models / evaluation: Leave-one-subject-out approach Training Testing Set Set Propositions: 1. Origin of data matters 2. Position matters 3. Self-Improvement matters Algorithms:

  • Classification-based:

SVM, ANN, Random Forest, k-NN, J48 Decision Tree

  • Threshold-based (params learned)
  • Supports deep learning,

but dataset is too small Data: MMSys UMA SisFall UniMiB #32 #10 #30 #23

slide-17
SLIDE 17

17

Hips Do Lie! A Position-Aware Mobile Fall Detection System

  • C. Krupitzer, T. Sztyler, J. Edinger, M. Breitbach,
  • H. Stuckenschmidt, and C. Becker

Experiment 1: Cross dataset analysis Experiment 2: Position-aware fall detection

Evaluation Overview

Experiment 3: Self-improving fall detection

slide-18
SLIDE 18

18

Experiment 1: Cross-Datasets Fall Detection

Hips Do Lie! A Position-Aware Mobile Fall Detection System

  • C. Krupitzer, T. Sztyler, J. Edinger, M. Breitbach,
  • H. Stuckenschmidt, and C. Becker

Trained across datasets Tested on Trained on same dataset

  • Proposition 1: Origin of data matters
  • The results indicate issues in labeling and setup of capturing data

0,52 0,68 0,63 0,48 0,73 0,71 0,45 0,54

0,00 0,20 0,40 0,60 0,80 MMSys UMA UniMiB SisFall F2-Scores ANN Trained across datasets Trained on same datasets

slide-19
SLIDE 19

19

Experiment 2: Position-Aware Fall Detection

Hips Do Lie! A Position-Aware Mobile Fall Detection System

  • C. Krupitzer, T. Sztyler, J. Edinger, M. Breitbach,
  • H. Stuckenschmidt, and C. Becker

Trained on one position Tested on all positions

0,82 0,80 0,61 0,69 0,65 0,73 0,80 0,76 0,57 0,00 0,20 0,40 0,60 0,80 Tested on Chest Tested on Waist Tested on Hip F1-Scores Model trained

  • n Chest

Model trained

  • n Hip

Model trained

  • n Waist
  • Proposition 2: Position matters
  • The results indicate that hips performed worst
slide-20
SLIDE 20

20

Experiment 3: Self-Improving Fall Detection

Hips Do Lie! A position-aware mobile fall detection system

  • C. Krupitzer, T. Sztyler, J. Edinger, M. Breitbach,
  • H. Stuckenschmidt and C. Becker

Falls Raw Data Position

slide-21
SLIDE 21

21

Experiment 3: Self-Improving Fall Detection

Hips Do Lie! A position-aware mobile fall detection system

  • C. Krupitzer, T. Sztyler, J. Edinger, M. Breitbach,
  • H. Stuckenschmidt and C. Becker

Falls Raw Data Position

slide-22
SLIDE 22

22

Experiment 3: Self-Improving Fall Detection

Hips Do Lie! A position-aware mobile fall detection system

  • C. Krupitzer, T. Sztyler, J. Edinger, M. Breitbach,
  • H. Stuckenschmidt and C. Becker

Falls Raw Data Position

slide-23
SLIDE 23

23

Experiment 3: Self-Improving Fall Detection

Hips Do Lie! A position-aware mobile fall detection system

  • C. Krupitzer, T. Sztyler, J. Edinger, M. Breitbach,
  • H. Stuckenschmidt and C. Becker

Falls Raw Data Position

slide-24
SLIDE 24

24

Experiment 3: Self-Improving Fall Detection

Hips Do Lie! A position-aware mobile fall detection system

  • C. Krupitzer, T. Sztyler, J. Edinger, M. Breitbach,
  • H. Stuckenschmidt and C. Becker

Falls Raw Data Position

slide-25
SLIDE 25

25

Experiment 3: Self-Improving Fall Detection

Hips Do Lie! A position-aware mobile fall detection system

  • C. Krupitzer, T. Sztyler, J. Edinger, M. Breitbach,
  • H. Stuckenschmidt and C. Becker

Detected falls: ANN J48 KNN Random Forrest ALM Position Chest Hip Chest Hip (Ground truth) (Classified) XYZ acceleration data

Detected falls: ANN J48 KNN Random Forest Self-Improving

  • Recall: 0.93 - Any falls missed?
  • No: We missed fall windows, but for each fall we detected at least
  • ne window!
slide-26
SLIDE 26

26

Experiment 3: Self-Improving Fall Detection

Hips Do Lie! A position-aware mobile fall detection system

  • C. Krupitzer, T. Sztyler, J. Edinger, M. Breitbach,
  • H. Stuckenschmidt and C. Becker

Detected falls: ANN J48 KNN Random Forrest ALM Position Chest Hip Chest Hip (Ground truth) (Classified) XYZ acceleration data

  • Proposition 3: Self-improvement matters
  • Precision is improved  decreases false positives

Detected falls: ANN J48 KNN Random Forest Self-Improving

slide-27
SLIDE 27

27

Current State

Hips Do Lie! A Position-Aware Mobile Fall Detection System

  • C. Krupitzer, T. Sztyler, J. Edinger, M. Breitbach,
  • H. Stuckenschmidt, and C. Becker

Acc.

Z Y X

slide-28
SLIDE 28

28

Current State

Hips Do Lie! A Position-Aware Mobile Fall Detection System

  • C. Krupitzer, T. Sztyler, J. Edinger, M. Breitbach,
  • H. Stuckenschmidt, and C. Becker

Fall Alert

Acc.

Z Y X

J48 Dec. Tree Random Forest SVM kNN ANN Thres- hold

slide-29
SLIDE 29

29

Future Work

Hips Do Lie! A Position-Aware Mobile Fall Detection System

  • C. Krupitzer, T. Sztyler, J. Edinger, M. Breitbach,
  • H. Stuckenschmidt, and C. Becker

Acc.

Z Y X

Fall Alert

Gyro Magn.

  • 1. Extension:

Considering

  • ther sensors

J48 Dec. Tree Random Forest SVM kNN ANN Thres- hold

slide-30
SLIDE 30

30

Future Work

Hips Do Lie! A Position-Aware Mobile Fall Detection System

  • C. Krupitzer, T. Sztyler, J. Edinger, M. Breitbach,
  • H. Stuckenschmidt, and C. Becker

Fall Alert

  • 1. Extension:

Considering

  • ther sensors
  • 2. Extension:

Cross-positional sensor fusion

Acc.

Z Y X

Gyro Magn.

J48 Dec. Tree Random Forest SVM kNN ANN Thres- hold

slide-31
SLIDE 31

31

Future Work

Hips Do Lie! A Position-Aware Mobile Fall Detection System

  • C. Krupitzer, T. Sztyler, J. Edinger, M. Breitbach,
  • H. Stuckenschmidt, and C. Becker

Fall Alert

  • 1. Extension:

Considering

  • ther sensors

N C F I M L Tn T2 I P K E Q T1

  • 3. Extension:

Optimized algorithms

Acc.

Z Y X

Gyro Magn.

... J48 Dec. Tree Random Forest SVM kNN ANN Thres- hold

  • 2. Extension:

Cross-positional sensor fusion

slide-32
SLIDE 32

32

Future Work

Hips Do Lie! A Position-Aware Mobile Fall Detection System

  • C. Krupitzer, T. Sztyler, J. Edinger, M. Breitbach,
  • H. Stuckenschmidt, and C. Becker

Fall Alert

  • 1. Extension:

Considering

  • ther sensors

N C F I M L Tn T2 I P K E Q T1

  • 3. Extension:

Optimized algorithms

Acc.

Z Y X

Gyro Magn.

  • 4. Extension:

Intelligent fall alert

... J48 Dec. Tree Random Forest SVM kNN ANN Thres- hold

  • 2. Extension:

Cross-positional sensor fusion

slide-33
SLIDE 33

33

Future Work

Hips Do Lie! A Position-Aware Mobile Fall Detection System

  • C. Krupitzer, T. Sztyler, J. Edinger, M. Breitbach,
  • H. Stuckenschmidt, and C. Becker
  • 1. Extension:

Considering

  • ther sensors

N C M L I P

  • 3. Extension:

Optimized algorithms

Acc.

Z Y X

Gyro Magn.

  • 4. Extension:

Intelligent fall alert

  • 5. Extension:

Outlier detection Fall Alert

... J48 Dec. Tree Random Forest SVM kNN ANN Thres- hold

  • 2. Extension:

Cross-positional sensor fusion

slide-34
SLIDE 34

34

Big Picture

Hips Do Lie! A Position-Aware Mobile Fall Detection System

  • C. Krupitzer, T. Sztyler, J. Edinger, M. Breitbach,
  • H. Stuckenschmidt, and C. Becker

Your project could appear here...

Fall Detection Intelligent Transportation Systems [ICAC17] Smart Vacuum Cleaner [PerIoT17]

Self-Improvement for Pervasive Systems

slide-35
SLIDE 35

35

The End

Hips Do Lie! A Position-Aware Mobile Fall Detection System

  • C. Krupitzer, T. Sztyler, J. Edinger, M. Breitbach,
  • H. Stuckenschmidt, and C. Becker
  • Shakira: Hips don’t lie
  • We: Hips do lie