Key Results

• Textures can be modeled as linear, non-Gaussian, stationary random fields.
• Textures can be synthesized using causal / non-causal AR models.
• AR model parameters can be estimated accurately using HOS.

Power spectrum estimation techniques have wide ranging applications. The information that is contained in the power spectrum is essentially that which is present in the autocorrelation sequence. Although this would be a complete statistical description for Gaussian signals, many real world signals are essentially non-Gaussian. Higher-Order Spectra (HOS) contain additional information which help in characterizing non-Gaussian signals. The main advantages in using HOS are in

• Detecting deviations from Gaussianity
• Extracting non-minimum phase information
• Detecting and characterizing non-linearity

Basics of HOS                                               Texture Model Validation

Texture Synthesis                                      Model Parameter Estimation

Summary                                                             Bibliography

Click here for a copy of our presentation on HOS.

back to Elec532 home page
last updated 5th May 2000