An auditory system modeling in sound localization

1
An auditory system modeling in sound localization

• EE381K MDDSP
• May. 5 2005
• Yul Young Park

2
Previous Talk

• Sound Localization motivations
• virtual 3D sound, game, sonar, microphone array,
  objective sound quality evaluation, education,
  etc
• Goal estimate correct azimuth and elevation
  (distance -gt ignored)
• Conventional approach
  neural network NetiYoung,1992 vs
  probabilistic estimator LimDuda,1994,
  ChauDudda,1995, Martn,1995
• Suggested method improve neural network model by
  evolutionary computation Stanley,2002

3
example block diagram LimDuda,1994
4
(No Transcript)
5
Neural Network
- network output
- desired output
- mean square error of network

• gradient descent method
• decrease the error by adjusting

• training data, test data
• issues local minimum, generalization, data
  representation

6
Evolutionary Computation stanley,2002

• Local Minimum, partly Generalization
• Evolution on Node and connection genes
  population -gt evaluation (fitness) -gt reproduce
  (ranking/sampling/crossover or mutation/speciation
  )
• Innovation record historical marking -gt easy
  topology analysis

7
Where I am
HRTF
Cochlear Neural Signal Model
Neuroevolution
Pre-Processing

• Elevation -4090, 10 increment
• Azimuth 530 at each elevation
• Total 710 points max.

• gammatone filterbank half wave rectification
• 64 ch. along 500Hz 22.05kHz

• population size
• fitness evaluation

• ITD cross-correlation
• IID subtraction

• computation issues data representation
  (int/real, bipolar/unipolar), dataset size
  (resolution/time)