Search engines can easily read and process text, although not in natural language yet. They actually "love" text because it can easily help them make sense of the pages they are crawling. The challenge with images and videos is that search engines cannot "read" them, yet.
They rely on the tags, descriptions, and the title of the page to categorize and index images and videos.
A good idea to help search engines understand these media better is to encourage users to describe, instead of comment on them. The typical comments to a video are; "nice", "cool", "very bad" etc... Nothing descriptive of the image or video.
Implementing this would give you two main advantages:
1. Better indexing for your pages: since your page will be full of relevant keywords that actually describe your content, the search engines will find it much easier. This is especially beneficial for sites with huge content, in leveraging a very long tail of keywords. Someone looking for "dog with green shirt" will definitely land on your page if a user described it this way. Moreover, the user will be happy to have found exactly what he is looking for.
2. Rich user experience: having a synopsis of a video, or a creative description for an image will help the viewer in recognizing certain things they would have not recognized otherwise. A new angle, or a special detail would give the content a new perspective. This would especially helpful with videos, since many times you wait several minutes hoping it is the content you are looking for, and then discover it is not. This way users are prepared for what they will watch.
It is not as easy as it sounds (I have yet to test it and see for myself) especially if you have users already user to just drop simple comments.
A good idea might be to introduce descriptions with (and eventually replace) comments. Users who don't want to write descriptions can vote for each description. This way, the best description rises in position and gives new users the best description available.
The level of interaction with the content goes up when people try to describe, instead of just venting their impressions. The number of people participating will definitely be less, because naturally, it is more difficult to describe than to just comment. I think it is a risk worthy of taking, to improve the way people use your service and to differentiate it among competitors.