Everything is possible, but I think this particular request might be a bit complicated. It should be possible to get the image URL with Data Scraping (with small modification to the XML that handles the Data Scraping).
When you click the “Edit Data Definition” in the Data Scraping wizard, you will get scary looking XML. However, there is some logic to it. It is better visible if you inspect the element with your browser.
You can see here that I didn’t find anything by selecting the images in Data Scraping:
But, you can modify the xml to get your results, it is actually really easy. You need to change the attribute you want to get based on the one that contains the URL when you inspect the element in the browser. In this case, we don’t want
<extract>
<column attr="text" name="Column1" exact="1">
but rather
<extract>
<column attr="style" name="Column1" exact="1">
As you can see, this simple change gives you the URLs 
It doesn’t give all because the container for default (no image uploaded) has different structure, but that should not be an issue.
You can now clean it up a bit and use the URL to download the images. Please follow this thread for that:
I hope it helps