MLx Home | Contents | MLx menu | MLx buttons | Widgets | Index | What's New

PCA - Principal Component Analysis


Theory

An introduction to what PCA can do as far as images are concerned is in:

David S. Bright, Principal Component Analysis , Scatter Diagrams and Color Overlays for analysis of Compositional Maps, MICROBEAM ANALYSIS 1995, pp 403-4, VCH, NY.


Practise

Superconductor Precursor Example


Functions

   
 windows -> sb16 stack
 doit
 --
 print Eigenvalues
 Print P matrix
 Display P matrix
 Colorize P matrix
 --
 Reconstruct Data
 plot P matrix

  1. windows -> sb16 stack This is puts the images into a special gray level image stack so that 'doit' will be able to make a similar stack of the principal component (eigenvalue) images. This PCA routine (need references) mean centers the (integer) data. The PC images and reconstructed images are also mean centered - scale to byte before writing a file for exporting to Image or Photoshop.
  2. Doit Perform the principal component analysis.