Bit confusing your question. You talk about 'responsive' design, which means that images are rescaled all the time. It seems logical that if you embed text in a bitmap image that this will result in blurriness. If you use real text in the Captivate project, this will not be the case. Or you can use SVGs, starting from a high resolution PNG, they will remain crisp when rescaled because they are vector-based.
The recommendation to use images only in the size you need them in a Captivate project is valid for a NON-responsive project which you publish with a fixed resolution.