Introduction to Hyperspectral/Multivariate Image Analysis

Course Description

Intro to Hyperspectral/Multivariate Image Analysis (MIA) is designed to give the student practical experience. Before the course, students will be sent a precourse reading assignment covering some of the basic background and principles of MIA. The course will start with a brief review of principal components analysis (PCA) and partial least squares (PLS) regression and how they are used in image analysis. Additional topics to be covered included multivariate image regression, and preprocessing to capture textural information. Methods to mitigate the effects of background interference, e.g. clutter, will also be discussed. The course includes hands-on computer time for participants to work example problems using PLS_Toolbox and MIA_Toolbox, or Solo+MIA.

Prerequisites

MATLAB for Chemometricians, Linear Algebra for Chemometricians, and Chemometrics I--PCA, or equivalent experience.

Course Outline

  1. Intro to 3-way arrays
    Objects and Variables
    Example Applications
    Structure of Multivariate Images
    Comparison to other sources of 3-way data
  2. Practical Multivariate Image Analysis (MIA)
    Review of Principal Components Analysis
    Scores, loadings and projections
    Unusual samples, residuals and T^2
    Matricizing of images
    Scores images, loadings
    Overlays
    Score/score plots: density
    Links between scores space and the image plane
    Contrast enhancement
    Image SIMCA
  3. Multivariate Image Regression analysis (MIR)
    Review of regression: MLR/PCR/PLS
    Scores, loadings
    Image plane and score linking
    Cross validation for images
  4. Preprocessing
    Centering and scaling
    Smoothing and derivitizing
    Scatter correction
  5. Intro to texture analysis
    Finite Fourier Transform (FFT)
    SVD Spectrum
    Angle Measurement Technique (AMT)
    Kriging
    MIR using texture transforms
  6. Alternatives to PCA/PLS
    Multivariate Curve Resolution
    PARAFAC on series of images
    Classical Least Squares
    Positive Matrix Factorization
    Generalized Least Squares and decluttering
    Extended addition model
    Evolving Window Factor Analysis
    Target Factor Analysis