iType: Using Eye Gaze to Enhance Typing Privacy Zhenjiang Li 1 , Mo - - PowerPoint PPT Presentation

▶
itype using eye gaze to enhance typing privacy
SMART_READER_LITE
LIVE PREVIEW

iType: Using Eye Gaze to Enhance Typing Privacy Zhenjiang Li 1 , Mo - - PowerPoint PPT Presentation

May 16, 2023 19 likes •179 views

iType: Using Eye Gaze to Enhance Typing Privacy Zhenjiang Li 1 , Mo Li 2 , Prasant Mohapatra 3 , Jinsong Han 4 , Shuaiyu Chen 4 CityU 1 , NTU 2 , UC Davis 3 , XJTU 4 Wearables Accelerometers Gyroscope Ambient light sensor Hart rate

slide-1
SLIDE 1

iType: Using Eye Gaze to Enhance Typing Privacy

Zhenjiang Li1, Mo Li2, Prasant Mohapatra3, Jinsong Han4, Shuaiyu Chen4 CityU1, NTU2, UC Davis3, XJTU4

slide-2
SLIDE 2

Wearables

  • Accelerometers
  • Gyroscope
  • Ambient light sensor
  • Hart rate sensor
  • Magnetometer
  • GPS
  • …

Extend beyond timing  daily life, e.g., fitness, exercise, business, etc.

[1] https://www.iphones.ru/wp-content/uploads/2015/05/main.jpg

slide-3
SLIDE 3

Explicitly typing sensitive info.

  • Password
  • Personal data
  • Security code
  • ….

Continuously sense hand moves

  • Accelerometers
  • Gyroscope
  • ….

However

slide-4
SLIDE 4

Wait a moment …

  • Touch ID
  • But

Account login Security code POS terminal Call support

Explicit Textual-Input is unavoidable

slide-5
SLIDE 5

Our idea for protection

  • Eye gaze for input
  • Front camera
  • Secure
  • Back
  • A keyboard
  • Front
  • Difficult to distinguish
  • Keyboard layout may change

*****

slide-6
SLIDE 6

iType framework

iType

_

1 2 3 4 5 6 7 8 9

/ ←

Frame Selector Password Assembler

Keyboard Rearranger

Accelerometers

Gaze Engine

xxxx

Front Camera Video Stream

Gaze Tracker

iType Engine

Typing Error Corrector Joint Decoder Button Selector

Keystroke Detector

Virtual Button

Transitional Gaze Remover Group Centroid Estimator

Flying Button

Enhance Layer

  • 3. Noises from

device motions

  • 1. Unreliable mobile

gaze tracking

  • 2. Lack of true text-entry

value in error correction

slide-7
SLIDE 7

Unreliable mobile gaze tracking

  • Problem statement:

Gaze tracker training [2]:

[2] “ishadow: design of a wearable, real-time mobile gaze tracker”, in Proc. of ACM MobiSys, 2014.

Input: image Output: gaze coordinates

slide-8
SLIDE 8

Unreliable mobile gaze tracking

  • Problem statement:

For mobile devices:

Training

slide-9
SLIDE 9

Unreliable mobile gaze tracking

  • Accuracy we need?

Smartphone Tablet

Error (degree) Error (degree)

slide-10
SLIDE 10

Unreliable mobile gaze tracking

  • Observations

Unreliable tracking

(a) (b)

x y

x-axis y-axis

slide-11
SLIDE 11

Unreliable mobile gaze tracking

  • Formal description
  • Solution overview (n gaze points)

Less More

  • Min. samples to achieve

certain confidence? At least (1- alpha)

slide-12
SLIDE 12

Keystroke detection

  • When to start?
  • KL divergence

Different!

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

(a) G1 G2 G3 (b)

k = 4 k = 4

Window size w = 12

slide-13
SLIDE 13

Keystroke detection

  • When to start:
  • KL divergence
  • When to stop:
  • Approximation

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

(a) G1 G2 G3 (b)

k = 4 k = 4

Window size w = 12

slide-14
SLIDE 14

Other modules

  • Input error correction
  • Joint decoding
  • Frame selection
  • Sensor-assisted

1 2 3 4 5 6 7 8 9

/ ←

c1

1 2 3 4 5 6 7 8 9

/ ←

c2

8 5 7 3 1 4 6 9

/

2

←

c2 ,

(a) (b) (c)

c2 ,,

slide-15
SLIDE 15

Evaluation

  • Overall performance

Individual keystroke:

  • Accuracy
  • Static: 97%
  • Dynamic: 89%
  • Latency
  • Static: 2.0s
  • Dynamic: 2.6s
slide-16
SLIDE 16

Takeaways

  • 1. On-going trend
  • 2. Dual aspects
  • 3. Challenges for iType

a) More & powerful sensors a) Beneficial to usage b) Potential privacy issue a) Unreliable mobile gazing b) Unknown ground truth c) Device motions

iType

_

1 2 3 4 5 6 7 8 9

/ ←

xxxx