The convolution defined the image and kernel

•Sections 3.4 to 3.6 in the Gonzalez and Woods textbook

Technical Approach

Carry out these operations to see the result for yourself. Compare the original image and the result. The convolution will iterate over each pixel, compute the average in a neighborhood defined by the filter, then take that average as the value of the filtered image.

A problem arises when you reach the border pixels. What do you do when the filter calls for averaging pixels that fall outside the boundary of the image? Consider a pixel at the origin (upper left). The filter will attempt to fetch f (-1;-1) but this pixel does not exist.

Page 2 of 4

Page 3 of 4

Your tasks:

2.Discuss the differences between border handling methods. Which is better for the given 7 X 7 filter? Which is better for larger filter sizes? Which is better for smaller filter sizes?

Page 4 of 4