Digital Image Processing with C++

3Point Processing Transformations

In this chapter, we are interested in grayscale images. A d-dimensional image is considered as a function I:Ω⊂ℕd→[​[0,N−1]​], where N is the number of gray levels that pixel values can take.

The simplest way to process an image is to apply a pointwise operation

J(x,y)=T[I(x,y)],

where T acts directly on the image I, or on the histogram of I. Gonzalez and Woods [15] propose many examples of ad hoc transformations. In the following, we give an overview of some of them.

3.1 Image operations

3.1.1 Mathematical transformations

A natural idea is to define a mathematical transformation T on I. Any “classical” mathematical function can be applied, regarding its definition domain. The CImg < T > class has many pre-defined ...