Image Processing - Report #4

This part of the report deals with the various uses of histogram equalization. There are four options associated with the equalization :

1. Global equalization : where equalization is performed on the entire image.
2. Selected area equalization : where an area specified by the user is equalized while the rest of the image remains the same.
3. Local equalization : Image is divided into 4 areas and equalization is done individually on the four areas and the combined result is displayed.
4. If the user is not satisifed with the result of the above operation, the image can be further divided and histogram equalization is done again.

Source code:

• Histogram - includes functions for histogram computation as well as equalization.
• Histogram Frame - draws the histogram to the screen.

Sample pictures:

HE of an area	Global HE
Local HE (4 areas)	Local HE (16 areas)

• Histogram before equalization
  
• Histogram after equalization
  

Analysis:
Histogram equalization results in increased contrast. As seen above, the image lax.gif has been significantly enhanced. Local equalization leads to a further improvement in the image. The features become more visible. However, local equalization may not be good in all cases, especially when the histogram is very narrow (the intensities are confined to a narrow band). The pictures of the heart show what happens when you apply global and local equalization to such an image. The image on the left is the original, in the middle is the globaly equalized image and on the right is the locally equalized image.

Below is a picture of peppers where local equalization led to interesting result. Even though the final image is changed, it is neat to look at!!!!