Median Filtering

Building the Cantata Workspace

1. Read image $DIP/data/bridge-toledo.viff and display it using operators User defined and Display Image.
2. Corrupt the image with shot noise by using operator Shot Noise, and display the result.
3. Filter the corrupted images using median filters with different window sizes using operator Median .
4. Filter the corrupted image using mean filter of size 5 x 5 using the operators Constant and Linear Operators
5. Display each image after it has been filtered.

Exercises

• Perform the same experiment with different percentages of noise corruption.
• Measure the similarity between the original image and the filtered one using mean square root metric. Which of the used filteres shows the better performance? Is this metric a good measurement of the performance of the filter?