Smoothing

 

Home Up Publications Research Projects Source Code Contact

 

Back Next

Next: Edge Detection Up: The CMSC 426 Image Previous: Thresholding

 

Smoothing

Image smoothing operations attempt to eliminate abrupt edges or changes in pixel intensity in an image. It is often thought of as ``blurring'' as well. It is usually accomplished by setting a pixel's value to be some function of the values of the pixels in a small neighborhood surrounding the pixel. The function may be anything from a simple average of the neighborhood pixels to more sophisticated gaussians.

There are three smoothing filters (or functions) implemented in the CMSC 426 image workshop. The first is a simple mean (averaging) filter. This filter can be chosen in three sizes: 3 x 3, 5 x 5, and 7 x 7. These filters simple calculate the average pixel intensity over their respective neighborhoods. The second is filter type is a guassian filter. This filter also comes in the same three sizes. The guassian filters are approximations to the integral of a guassion function over the neighborhoods, which are then normalized to produce values in the correct intensity range. The exact filters we used are as follows:

 

  1. tabular29
  2. tabular32
  3. tabular35


 

Vasanth Philomin
Thu Jun 11 12:35:41 EDT 1998