XSLT can be used to transform your input XML into a different XML format (for example, XSL Formatting Objects) or into HTML. It can't "render images" directly, it can only generate output in a format that some other software (for example a web browser) is capable of rendering.
If your target environment is HTML, then start by sketching out the HTML you want to produce. Handling the image references is no different from any of the rest of the information - you're only manipulating text.
Author, XSLT 2.0 and XPath 2.0 Programmer\'s Reference