Wave to Me: User Identification Using Body Lengths and Natural Gestures, at CHI 2014
•Télécharger en tant que PPTX, PDF•
0 j'aime•322 vues
We introduce a body-based identification system that leverages individual differences in body segment lengths and hand waving gesture patterns. The system identifies users based on a two-second hand waving gesture captured by a Microsoft Kinect. To evaluate our system, we collected 8640 gesture measurements from 75 participants through two lab studies and a field study. In the first lab study, we evaluated the feasibility of our concept and basic properties of features to narrow down the design space. In the second lab study, our system achieved a 1% equal error rate in user identification among seven registered users after two weeks following initial registration. We also found that our system was robust even when lower body segments could not be measured because of occlusions. In the field study, our system achieved 0.5 to 1.6% equal error rates, demonstrating that the system also works well in ecologically valid situations. Lastly, throughout the studies, our participants were positive about the system.
Recommandé
Recommandé
Contenu connexe
Similaire à Wave to Me: User Identification Using Body Lengths and Natural Gestures, at CHI 2014
Similaire à Wave to Me: User Identification Using Body Lengths and Natural Gestures, at CHI 2014 (20)
Dernier
Dernier (20)
Wave to Me: User Identification Using Body Lengths and Natural Gestures, at CHI 2014
- 1. Wave to Me: User Identification Using Body Lengths and Natural Gestures Eiji Hayashi Manuel Maas Jason Hong Human-Computer Interaction Institute Carnegie Mellon University
- 6. Slick user identification with reasonable security
- 7. Gesture
- 8. EachDifferent Gesture for user
- 9. AllSame Gesture for user
- 11. 11 97% accurate
- 12. System Lab Study 1 (Basic Evals) Lab study 2 (Long-term Eval) Field Study
- 13. System Lab Study 1 (Basic Evals) Lab study 2 (Long-term Eval) Field Study
- 19. Kinect
- 20. Kinect
- 24. Kinect Feature Extraction Physiological 17 body segment lengths Behavioral 26 movement properties
- 29. Kinect Feature Extraction SVM Pre-Recorded Data User ID + Confidence
- 32. Errors False Acceptance Rate (FAR) Accept others as a registered user False Rejection Rate (FRR) Reject a registered user as others
- 33. Errors False Acceptance Rate (FAR) Accept others as a registered user False Rejection Rate (FRR) Reject a registered user as others Equal Error Rate (EER) FAR = FRR = EER
- 34. Errors False Acceptance Rate (FAR) Accept others as a registered user False Rejection Rate (FRR) Reject a registered user as others Equal Error Rate (EER) FAR = FRR = EER Accuracy = 1 – 2 x EER
- 35. Assumption There are 7 registered users in our system
- 36. Assumption There are 7 registered users in our system Make comparison among studies easy
- 37. Assumption There are 7 registered users in our system Make comparison among studies easy Be reasonable for home use
- 38. Assumption There are 7 registered users in our system Choose 10,000 combination of 7 participants Calculate EER over them
- 39. System Lab Study 1 (Basic Evals) Lab study 2 (Long-term Eval) Field Study
- 40. Gestures Hand Waving Come-Over One Hand Raised Making a Phone Call
- 41. Data Collection Hand Waving Come-Over One Hand Raised Phone Call Gesture
- 42. Data Collection Hand Waving Come-Over One Hand Raised Phone Call Standing Sitting Gesture Posture
- 43. Data Collection Hand Waving Come-Over One Hand Raised Phone Call Standing Sitting 1st Day 3 days later Gesture Posture Session
- 44. Data Collection Hand Waving Come-Over One Hand Raised Phone Call Standing Sitting 1st Day 3 days later Gesture Posture Session 10
- 45. Data Collection Hand Waving Come-Over One Hand Raised Phone Call Standing Sitting 1st Day 3 days later Gesture Posture Session 10 160 / participants
- 46. Participants 36 participants 14 males / 22 females 19 – 64 years old 168cm (SD=10.2) 78.0 kg (SD=22.0)
- 47. Gestures Hand Waving Come-Over One Hand Raised Making a Phone Call
- 48. Using Either Gesture or Lengths Same day & posture 3 days later Different Posture Same day & posture 3 days later Different Posture Gesture Body Lengths EER [%]
- 49. Using Either Gesture or Lengths Same day & posture 3 days later Different Posture Same day & posture 3 days later Different Posture Gesture Body Lengths EER [%] 2.1% 0.5%
- 50. Using Either Gesture or Lengths Same day & posture 3 days later Different Posture Same day & posture 3 days later Different Posture Gesture Body Lengths EER [%] 11.8% 19.8%
- 51. Using Either Gesture or Lengths Same day & posture 3 days later Different Posture Same day & posture 3 days later Different Posture Gesture Body Lengths EER [%] 10.0% 41.5%
- 52. Using Both Gesture and Lengths 3 days later 3 days later Gesture Body Lengths EER [%] Both 3 days later (Standing) 3 days later (Sitting) 4.3% 6.2%
- 53. System Lab Study 1 (Basic Evals) Lab study 2 (Long-term Eval) Field Study
- 54. Data Collection Hand Waving Standing Sitting 1st Day 3 days later 1 week later 2 weeks later Gesture Posture Session 10 80 / participants
- 55. Participants 27 participants 20 males / 7 females 19 – 62 years old 173cm (SD=9.8) 75.1 kg (SD=21.1)
- 58. Long term StabilityEER[%] Days Stable after the 3rd session
- 59. Training with 2 sessionsEER[%] Days
- 60. Training with 2 sessionsEER[%] Days Sitting Standing
- 61. Training with 2 sessionsEER[%] Days EER < 1%
- 62. System Lab Study 1 (Basic Evals) Lab study 2 (Long-term Eval) Field Study
- 63. Does it Work at Homes? • Collected data at participants’ living rooms • Placed a Kinect on a TV • Asked participant to behave as usual – Stand where you feel reasonable – Sit as you normally do in a living room
- 64. Participants 12 participants (5 house hold) 5 males / 7 females 18 – 42 years old 159.5cm (SD=11.4) 56.9 kg (SD=6.9)
- 65. It Worked! EER [%] Standing Sitting Standing Sitting Lab Study 2 Field Study
- 66. Implication Recognizing a gesture Recognizing a gesture AND a user’s identity
- 67. • Natural gestures + body lengths • 2 seconds of hand waving gesture Conclusion Gesture Body Lengths Proposed Scheme EER [%]
- 68. Wave to Me: User Identification Using Body Lengths and Natural Gestures Eiji Hayashi ehayashi@cs.cmu.edu www.cs.cmu.edu/~ehayashi/ Human-Computer Interaction Institute Carnegie Mellon University
- 69. Backup slides
- 70. # of Registered UsersEER[%] # of Registered Users
- 71. # of Registered UsersEER[%] # of Registered Users 2.8% 2.3% N=25
- 72. Open Questions Getting worse constantly? Training with two sessions?
- 73. Data Collection Hand Waving Standing Sitting 1st Session 2nd Session 3rd Session Gesture Posture Session 10 60 / participants
- 74. Yet Another Open Question Does it actually work at home?