Construct a dynamic error cancelation filter, using computer vision

• Useful for correcting EEG signal errors
• Should be used in conjuction with other filters
  • High pass
  • Low pass
  • 60 Hz removal

Remove signals due to music infiltration in the EEG signal

• Requires classification of music
• Then removal of associated music interferance

Record a video/eeg of a user as they have various states, over 5 minutes

• 0 - 20 seconds: "Blank Face"
• 21 - 40 seconds: Blinking
• 41 - 60 seconds: Right Wink
• 61 - 80 seconds: Left Wink
• 81 - 100 seconds: Fake Talk
• 101 - 120 seconds: Furrow Brow
• 121 - 140 seconds: Smile
• 141 - 160 seconds: Squint
• 161 - 180 seconds: Scratch Right Ear
• 181 - 200 seconds: Scratch Left Ear
• 201 - 220 seconds: Scratch Right Face
• 221 - 240 seconds: Scratch Left Face
• 241 - 260 seconds: Close Eyes& Count to 15
• 261 - 280 seconds: Act Like Chewing Gum
• 281 - 300 seconds: View Flashing Box @ 5 Hz

Track face and normalize

• Accomplished by tracking the face
• Constructing a retangle around the face
• Construct a grid inside the rectangle
• Determine movement inside each grid selection (CV algorithm)

Compare EEG signals with grid movement

• Associating each EEG signal with a given grid movement
• Accomplished by using feature recognition
• Specifically, search for alpha, theta, beta waves