Yeah.  Using RepositoryAccess, you'd want to do something like:

PageInfo page = // however you get your page

using (WordLocationsReader words = page.ReadLocationsPagePart())
{
    while (!words.EndOfStream)
    {
        LfRectangle rect = words.Read();
        // if this is a word you want, do what you need to with the location
    }
}

To unpack a little bit, the location data is stored in the "locations" page part, and so ReadLocationsPagePart() is used to return a WordLocationsReader, which in turn will produce a sequence of LfRectangles, one for each word in the text on the page (assuming that the text was generated by OCR).  So you'll need to iterate through the list of rectangles until you find the ones corresponding to the parts of the text you care about, then do whatever you need to do with them.