Chapter 3: Image processing

This chapter describes the steps to developing a typical application with the MIL Image Processing module.

• Image processing overview
• • MIL and image processing
• Steps to performing a typical application
• • A typical application
  • How to encode these steps
• Image quality
• • Techniques to improve images
• Averaging an input sequence
• Denoise using spatial filtering and area open and close operations
• • Low-pass spatial linear filters
  • Rank filters
  • Area open and area close
• Erosion and dilation
• • Basic erosion
  • Basic dilation
  • An example
• Opening and closing
• Basic geometric transforms
• Image statistics
• • Generating a histogram
  • Calculating general statistics
  • Considering pixel values as unit vectors and calculating angular statistics
  • • Coherence
    • Dominant angle
    • Dominant orientation
  • Calculating statistics across multiple images
  • Finding the image extremes
  • Locating events
  • Counting image differences
  • Projecting an image to one dimension
• Thresholding your images
• • Binarizing
  • • Determining the threshold value from the histogram
    • Specifying the threshold value(s) manually
  • Clipping
• Adaptive binarizing
• • Adaptive binarize contexts
  • Controlling an adaptive binarize context
  • • Threshold with Niblack
    • Threshold with local mean
    • Threshold with Bernsen
    • Threshold with pseudomedian
    • Hysteresis
  • Controlling an adaptive binarize context that uses seeds
  • • Threshold with geodesic reconstruction
    • Threshold with leveling
    • Threshold with toggling
• Histogram equalization
• • Adaptive histogram equalization
• Enhancing and detecting edges
• • Edge enhancement
  • • Sharpening edge enhancement
  • Edge detection
  • • Horizontal and vertical edge detection
    • Laplacian-based edge detection
    • Gradient-based edge detection
• Arithmetic with images
• • Combining images
  • Mapping an image
• Distance transform
• • City Block transform
  • Chessboard transform
  • Chamfer 3-4 transform
• Labeling