Image Processing Operations

Documentation external to the source code is provided for the following operations:

• Rasterop (a.k.a. Bitblt)
  • What is rasterop?
  • What else can you do with rasterop?
  • How is rasterop implemented efficiently?
  • What special cases have been implemented?
  • What is the calling sequence for rasterops?
• Binary Morphology
  • Introduction
  • References and meta-references
  • What is binary morphology?
    • And there is more to morphology than translationally-invariant, increasing primitives
  • What about boundary conditions?
  • The rasterop implementation of binary morphology
  • The destination word accumulation implementation of binary morphology
    • Auto-generated dwa code interfaces
  • Separable atomic operations with block Sels, Sel decomposition, sequence interpreters and simplicity of use
    • Atomic separable functions
    • Interpreters for sequences of atomic operations
  • Summary of available implementations
    • Full image rasterops
    • Destination word accumulation
    • Sequence interpreters
    • Performance comparisons
  • Binary morphology and cellular automata
• Grayscale Morphology
  • What is grayscale morphology?
  • Implementation using the van Herk/Gil-Werman algorithm
  • The Tophat and H-dome transforms
  • Generalization: rank order filters
    • Definition
    • How is a brick rank order filter implemented efficiently?
    • Points of interest in the implementation
• Fast Convolution
  • What is a convolution?
  • Grayscale convolution using an accumulator
  • Binary rank-order and median filter using an accumulator
• Image Scaling
  • Overview of image scaling
    • What types of image scaling are useful?
  • Grayscale Image Scaling
    • How do you scale a grayscale image?
      • Sampling
      • Area mapping (or area averaging) and lowpass filtering
      • Mip-mapping
      • Min-max
      • Linear interpolation
    • Implementation of linear interpolation scaling
    • Considerations of efficiency for linear interpolation scaling
  • Binary Image Scaling
    • Arbitrary binary rescaling
    • Special case: power-of-2 binary expansion
    • Special case: 2x binary reduction
  • Binary to Gray Downscaling (“scale-to-gray”)
  • Gray to binary upscaling (“scale-to-binary”)
  • Fast scaling with depth change
  • Summary of scaling functions in leptonica
• Image Rotation
  • How do you rotate an image by an arbitrary angle?
  • Rotation by Shear
  • Rotation by Area Mapping
  • Fast rotation by Area Mapping
  • Special rotations by 90, 180 or 270 degrees
    • 180 degree rotation
    • 90 degree rotation
• Affine Transformations (and cousins)
  • What is an affine transformation?
  • How are affine transformations implemented?
    • Sequential affine transform
  • Some related transforms
    • Conformal affine transform
    • Projective transform
    • Bilinear transform
  • Shearing an image
• Seed Filling and Connected Components
  • What is the seed fill operation?
  • Seed filling to identify connected components
  • Seed filling using binary morphology
  • Seed filling using iterative raster-ordered filling
  • Seed filling extended to grayscale images (grayscale reconstruction)
  • Generation of a distance function within connected components
  • Labelling pixels using distance function and grayscale reconstruction
  • Watershed transform on seeded grayscale images
• Grayscale and Color Image Enhancement
  • What do we mean by grayscale and color image enhancement?
    • What other types of enhancement operations exist?
  • Gamma correction
  • Contrast enhancement
  • Unsharp masking (edge enhancement)
  • Image smoothing
• Grayscale Mapping and Binarization
  • Locally adapted grayscale mapping
  • Locally adapted binarization
    • Sauvola binarization
    • Local background normalization
    • Local background normalization with Otsu
    • Local background normalization with two thresholds
  • Global binarization
• Connectivity-preserving Thinning