In my understanding, the key point is that both the Imaging or Canvas APIs are still considered experimental.
In particular, their APIs aren't frozen, and until we finalize an API that we want to commit to maintaining, neither API has been publically released. Also, there is enough overlap between the two APIs (the Canvas API arguably offers a superset of the Image API's functionality) that it seems silly to try to launch both, and whilst the Image API (which came first) is more or less complete, the (currently partially implemented) Canvas API is probably the better long-term candidate, given that the <canvas> that inspired it is part of the HTML5 specification. Unfortunately, I don't know anyone who is actively working on this. But if you were to pick one API to contribute code to, I would recommend the Canvas API.
