Hello,
I need to compare the similarity between curves. Two curves X and Y are said to be similar if they satisfy any one of the following (or combinations):
1/ Curve X and Y obey some translation and they have the same overall shape e.g curve X starts at coordinate (0,10) and curve Y starts at (0,21) but they each have the same overall shape and frequency.

 2/ Curve X is an expansion/enlargement of Curve Y along the x axis but their shapes are somewhat preserved e.g. sine(f) and sine(2f) are similar.

3/ Curve X is an expansion/enlargement of Curve Y aong the Y axis e.g. 1.23*sine(f) and 345*sine(f) i.e. different amplitudes but they are still similar.

It could also be a combination of the above. The example above uses sine waves but my waves are much more complex than this with loads of frequency contents.

Does anyone have a brief idea how I could achieve this? In the first instance i.e. 1/ above I could compute the frequency spectrum and determine if both X and Y have the same frequency components but the problem is that I could have a combinations all three and the complexity keeps increasing.
Thanks!

"Mat Lung" wrote in message <ievrav$jmd$1@fred.mathworks.com>...
> Hello,
> I need to compare the similarity between curves. Two curves X and Y are said to be similar if they satisfy any one of the following (or combinations):
> 1/ Curve X and Y obey some translation and they have the same overall shape e.g curve X starts at coordinate (0,10) and curve Y starts at (0,21) but they each have the same overall shape and frequency.
> 
>  2/ Curve X is an expansion/enlargement of Curve Y along the x axis but their shapes are somewhat preserved e.g. sine(f) and sine(2f) are similar.
> 
> 3/ Curve X is an expansion/enlargement of Curve Y aong the Y axis e.g. 1.23*sine(f) and 345*sine(f) i.e. different amplitudes but they are still similar.
> 
> It could also be a combination of the above. The example above uses sine waves but my waves are much more complex than this with loads of frequency contents.
> 
> Does anyone have a brief idea how I could achieve this? In the first instance i.e. 1/ above I could compute the frequency spectrum and determine if both X and Y have the same frequency components but the problem is that I could have a combinations all three and the complexity keeps increasing.
> Thanks!