I'd like to have the tex, to translate the letters to blocks before doing image processing, but it's almost as good to simply convert it to a binary mask, and t

I'd like to have the tex, to translate the letters to blocks before doing image processing, but it's almost as good to simply convert it to a binary mask, and then "dilate" or "grow" the pixels. I think it makes it a simpler problem both in processing time and conceptually. Not sure about how I'd decide, identify what vertical lines to pick out, there are a lot of choices for that.