For each recognition area you can use several image processing functions so that the recognition engine get a clean image and can give a reliable result. You can use a large number of advanced functions but an inappropriate use instead that increase recognition rate can decrease it, so you have to decide the function to use carefully and only if required.
You have to remember that pre-processing functions work only on the selected area and that they are not permanent: the original image is not modified unless you don't use some custom script function to do it.
Thanks to debug feature available in Recogniform Reader you can check visually the result of pre-processin in the testing phase, so that you can be sure of your choices.
The main functions you can use are: Deskek, Despeckle, Lines Removal, Box Removal, Streak Removal, Thinning, Thicking,.
You have to use this functions carefully: as example activating the deskeckle function on a field with dot-matrix printed text could eliminates all the characters dots and not only the speckle, so activating the line removal function on a barcode area could delete all the bars of the code !
But using the box removal function on a field with handwritten boxed text is essential to get the data recognition, so that to use togheter the thinning and thicking function could be very important to normalize the pen thickness of handwritten fields filled with pencil or stencil.