Trinetra Project

1
Trinetra Project
Alternative Pointing Device with Hidden Markov
Model
CSE 19991253 Seung hyun, Yang
2
Problems

• Present pointing device Mouse
• Occupies much place
• Needs ground
• Future pointing device EEG BCI
• Difficult to obtain data
• Poor reliability of output

3
Trinetra?

• Distance measuring to 3 directions
• Training with the measured data
• Classification the test data set

HMM, FHMM
Recognition
Pattern Classification
4
Pattern Classification

• Direction
• Expectation angle of the likelihood of each 3
  direction (with addition of vector)
• Click (down)
• Click (up)

Hidden Markov Model

• Assume that finger doesnt move in parallel along
  a wall.
• Optional implementation of moving distance

5
Data Gathering

• Senscomp. SE-600-I
• Ultrasonic sensor
• Measure the distance to an object
• Three directional measuring
• Parallel Port 11-13
• (input)

6
Data Gathering
7
Plan

• Prepare The Equipments (H/W)
• 6, Nov.
• Data gathering part
• 20, Nov.
• Implementation of classification
• 4, Dec.