I've been in that place before and this is what I have found.
First, it's hard to be creative in a vacuum. You need to be consuming media, even if it's not all games. Play games you haven't played before, read game design blogs, try to break down successful games' designs, figure out what works about them and why. Watch films, look at art & photography, read books, doodle, listen to music (that's a big one for me!), even if not all of it is game related. Interests outside of games can help you bring in ideas that are new to the world of games, and make more interesting games as a result. That being said, I think if you hardly play games at all (like I was for a while) that impacts your creativity/interest in that area.
I also get on weird compulsive binges that have nothing to do with game development (I spent about a month recently reading about woodworking
) and it's hard to ignore. Sometimes it works better to just indulge that interest for a while until the desire goes away.
It might help to try and think of different angles to approach your design from, like different lenses or perspectives. This book enumerates something like a hundred different lenses to look at design through: http://www.amazon.com/The-Art-Game-Design-lenses/dp/0123694965
You can probably just look through the table of contents to get an idea of some avenues to think about.
There are also creativity-enhancing devices, like brainstorming and Oblique Strategies ( http://www.rtqe.net/ObliqueStrategies/
) or the Video Game Name Generator ( http://videogamena.me/
). Brainstorming is hard for me because I think as I have gotten older I tend to filter myself much more, shooting down 'absurd' ideas before I fully consider them, but it's probably even more beneficial to practice if that's the case with you too.
If you're lucky like I occasionally am, you will play a video game in your dreams and be able to remember some interesting details.
Also be aware that some genres might well be described as mined out, in the sense that you're unlikely to invent a mechanic that hasn't somehow been represented in some other game in the past. Genres that have been around since the 80's, especially, are prone to this. I don't think that's necessarily an argument against making games in that genre, but it makes it more difficult. What you can do is assemble them into a new and never-before-seen whole, not invent everything new wholesale. Just be careful, as I've found you can't just mash two interesting games together into one, because usually the games are intricately optimized themselves, and in the space between them there may not be as many optimal peaks.
EDIT -- adding a couple more thoughts:
I've used randomness as a source of visual inspiration before. Stare at something with random texture (popcorn ceilings, rugs, etc.) and see if patterns emerge. It's pretty easy to see faces or animals in these things since our brain is wired to find those sorts of patterns.
I also read something once about Apple's design process -- one person makes 10 distinct designs for something. Usually the first 2 or 3 are easy and by the time you get to 10 it's really damn hard. At the end you pick the best one. It can be pretty hard coming up with 10 really different variations of some basic idea, so by the time you are on 7, 8, 9, etc. you are usually in pretty weird, unexplored territory, which is what you want.