Fractal video compression seems like an obvious thing to investigate.
I expect it has been explored and is not commonly used for good reasons.
Why? Computationally intensive? Too sensitive to noise?
the major problem of fractal image compression is the time effenciency. The
same problem arises with fractal video coding. Although there are a lot of
algorithms which can speed up the encoding process, the best fractal
encoders still need alot of time (minutes for one image).
If you are intersted in further details, just check this out:
Here you'll find alot of information about fractal image compression as
well as papers about fractal video compression.
> I expect it has been explored and is not commonly used for good reasons.
> Why? Computationally intensive? Too sensitive to noise?
Yes, it is possible in 500 lines of C code ;-). Although compression
run fine on PII/350 or a $100 specialized board. If anyone is interested
By the way: what if I used fractal mechanisms on 1-D output of 512-point
FFT on voice data? (it might be useful as the harmonics could be
described easily as self-similar fractal parts of main signal). Has
anybody tried this way?