I haven't used the default image/caption approach, which might be the "correct" answer to your question … but pending a better answer, I'd deal with your requirement like this:
- set up a style for images and a style for captions
- inserting each image in an empty paragraph with the image style
- add a caption after each image I want to include in the ToC, using the caption style
- add the caption style to the ToC definition
In general, styles are the key to customising tables of contents.