Grey Level Processing

Grey level processing operations involve transformation and manipulations with intensity of image pixels. The core set of operations include:

Shading Correction:

Image defects that are caused by uneven illumination or artifacts in the imaging path must be taken into account during the image processing. Shading Correction is used when a large portion of an image is darker or lighter than the rest of the image due to, for example, bulb misalignment or use poor optics in the system.

Shading correction may be performed using different correction methods

  • Using Reference Image with baseline shading
  • Rank Leveling with
  • Polynomial fitting of nearest-neighbor pixels

Pixel Point Processing:

Pixel Pointing operations are the class of image enhancements that do not alter the relationship of pixels of their neighbors. This class of algorithms uses a type of transfer function to translate the original grey levels into new grey levels, usually called a look-up table, or LUT.

In these transformations the pixel intensity value changed to new value based on original value

Histogram Manipulations:

If the signal and contrast mechanism are suitably selected, the peaks in the histograms correspond to phases in the structure.

 

 

Look-Up Table

Structure within nodules not apparent

Structure within nodules clealy

Inverse Scaling

Pseudo-color Look Up Table

Pseudo-Coloring

Neighborhood Kernel Processing

Neighborhood Kernel processing is a class of operations that translates individual pixels based on surrounding pixels. The concept of using a kernel or two-dimensional array of numeric operators provides a wide range of image enhancement.

Neighborhood Kernel Processing includes rank-order, Gaussian, Laplosian and averaging filters.

  • Rank filter
  • Transformations
  • Averaging
  • Gaussian
  • Laplacian
  • Many others

Rank Order Processing

Median Filter

Is an example of a rank-order filter. Median filter determines the median, or 50%, value of a selected kernel and replaces the central value with the median value.

3x3 kernal

32

56

75

25

47

60

95

77

80

 

rank intensities in ascending order

25        32        47        56        60        75        77        80        95

median value

32

56

75

25

60

60

95

77

80

 

replace central pixel value with median value