Pseudo-Inverse Filtering

Building the Cantata Workspace

1. Get the float type blurred image $DIP/data/lenina-blur-float.viff and convert its pixel type to unsigned byte using operators User defined and Convert Type operators. Display the blurred image.
2. Subtract the float type image from the unsigned byte image and display it.
3. Use the same steps of the building the workspace for inverse filtering lesson to find the restored image.
4. Implement the pseudo-inverse filter by performing a threshold on the magnitude of the DFT of the kernel filter. Then multiply it to the ideal inverse filtering. Find the restored image for this case.
5. Use operator Inverse Filter to implement directly pseudo-inverse filtering.

Exercises

• What happens if the threshold value used in the Inverse Filter is zero?
• Modify the filter threshold value in Inverse Filter.