Notice: To execute the interactive demo, you need to download the ipynb file to your computer. Please refer to Demo1a to learn how to install and use ipython notebook.

Section 1
  • Demo1a – Optical Illusion. Link to Michael Bach’s webpage
  • This webpage contains demos of many beautiful optical illusions and visual phenomena. Professor Bach gives detailed descriptions of these phenomena from a theoretical perspective which is similar to the viewpoint expressed in this chapter. He states that “I view these phenomena as highlighting particular good adaptations of our visual system to experience with standard viewing situations. These experiences are based on normal visual experiences, and thus under unusual contexts can lead to inappropriate interpretations of a visual scene (=Bayesian interpretation of perception).” We particularly draw attention to: (i) Hidden Figures, (ii) Rotating Face Masks, (iii) Ames Window, (iii) Neon Color Spreading, (iv) Dress Code Enigma, (v) Adelson’s “Checker-Shadow” Illusion, and (vi) Biological Motion.
  • Demo1b: IPython Notebook Basics. Link to Nature website
  • This demo links to a Nature article about how to use ipython notebook for research purpose. IPython notebook is needed for the interactive demos in the following sections. If you don't have experience with iPython Notebook, in this article you can try the provided interactive demo to get familiar with the usage of this tool.
    Section 2
  • Demo2a: Convolution Demo. | View | Download
    This demo introduces linear filters and convolution.
  • Demo2b: Gabor Demo. | View | Download
    This demo introduces Gabor filters.
  • Demo2c: PCA and Oja’s rule. | View | Download
    This demo illustrates Principal Component Analysis, and Oja’s rule.
  • Section 3
  • Demo3a: Natural Image Statistics. | View | Download
  • Demo3b: Statistical Edge Detection. | View | Download
    This demo illustrates decision theory using edge detection as an example.
  • Section 4
  • Demo4a: Gibbs Sampling. | View | Download
    This demo illustrates Gibbs sampling which serves as a simplified model of stochastic neurons.
  • Demo4b: Mean Field Theory. | View | Download
    This demo describes mean field theory which relates to deterministic neural network models and Hopfield networks.
  • Demo4c: Hopfield Network for Binocular Stereo. | View | Download
    This demo applies a Hopfield network (mean field theory) to binocular stereo.
  • Section 5
  • Demo5: Cue Combination. | View | Download
    This demo shows how cues combine weighted by their uncertainty as their variances change.