Stroke filters are not a standard idea for image processing. Instead, it is a term created for the document to which you are attached, where they determine what they mean by the impact filter (section 2), and how you can implement one computational one (section 4).
Because of this, you are unlikely to find implementation in the standard toolbox; although you may want to look around for those who published the implementation, or contact the authors.
Basically, they say that you can identify the text in the image based on the typical properties of the text, in particular that the text has many structures, like strokes, with a certain spatial distribution.
source share