![]() We compute the volume (integral) as a sum. In our discrete finite case, we represent our 2D functions as matrices of values. We will also call it "radius" in the text below. Our gaussian function has an integral 1 (volume under surface) and is uniquely defined by one parameter $\sigma$ called standard deviation. The Gaussian blur of a 2D function can be defined as a convolution of that function with 2D Gaussian function. The second function $g$ is sometimes called "weight", since it determines, how much of $f$ will get into the result The convolution of two 2D functions $f$ and $g$ is defined as the volume of product of $f$ and "shifted" $g$. I have implemented this code into Photopea under Filter - Blur - Gaussian Blur. To get motivated, take a glance at the results. I am going to describe it a little better and add some mathematics. Presented ideas are very simple and I don't know who is the original author. My solution is based on Fast image convolutions by Wojciech Jarosz. After hours of struggling and browsing the internet, I finally found the best solution. I needed a really fast Gaussian blur for one of my projects.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |