Advanced smooth filter


This application is a usage example of the pno "Convolution filter" which provides the possibility to apply different smooth filters onto a selected image (Exponential, Gaussian, ...).

It uses an editable matrix that can be filled by user to create specifical filters. At this time, the matrix coefficients can't be negative. You can change the pno code if you want to include this feature (for edge detection as example).

The matrix is quite flexible as its coefficients can be integer numbers or fixed point numbers. The fixed point numbers ensure best accuracy but increase time computing.

The integer division is not included in my mathematic library (this first version is not optimized and can be improved) but instead I use the one coming with the standard library (no -nostdlib flag in the make file).