Error Concealment | JPEG | MPEG-4 | Wavelet | YUV

Hi guys,

I am currently doing a project about the multiple description coding for a image using
correlating transform. In one of the IEEE journals, the author describes the encoder as 

“The coder is based on a modification of a standard JPEG coder. In our baseline encoder, each
block of samples is transformed using an 8*8 discrete cosine transform (DCT). These
coefficients are ordered to have decreasing variance.”

I have done the DCT part, but can’t figure out how to rearrange these coefficients in the
orders of decreasing variance. I thought the variance is only meaningful when it refers to a
set of random variables and the value of variance will not be affected by the ordering of
variables.

Could anyone help me with this problem? Thank you very much.

(You need to be a member of imagedsp -- send a blank email to imagedsp-subscribe@yahoogroups.com )

Hi,

Probably the author is using the term 'variance' for the variance of 
interblock coefficients (and not of intrablock coefficients). For 
example, if the variance of n'th DCT coeff (taking its values from 
different blocks of 8x8 size) is greater than the same of m'th coeff, it 
should come before it.
abhishek.
f...@hotmail.com wrote:

>Hi guys,
>
>I am currently doing a project about the multiple description coding for a image using
correlating transform. In one of the IEEE journals, the author describes the encoder as 
>
>“The coder is based on a modification of a standard JPEG coder. In our baseline encoder,
each block of samples is transformed using an 8*8 discrete cosine transform (DCT). These
coefficients are ordered to have decreasing variance.”
>
>I have done the DCT part, but can’t figure out how to rearrange these coefficients in the
orders of decreasing variance. I thought the variance is only meaningful when it refers to a
set of random variables and the value of variance will not be affected by the ordering of
variables.
>
>Could anyone help me with this problem? Thank you very much.
>

(You need to be a member of imagedsp -- send a blank email to imagedsp-subscribe@yahoogroups.com )

You haven't given the whole context, are the coeff's with smaller variance 
coded differently, discarded or something similar? Of course the variance is 
not affected by ordering, but the ordering gives you an insight on the block 
content, and enables you to act accordingly... and that's probably where the 
real "smarts" of the encoder will be.

Kind personal regards,

Nikola Gedelovski

Abhishek KUMAR wrote:
> Hi,
>
> Probably the author is using the term 'variance' for the variance of
> interblock coefficients (and not of intrablock coefficients). For
> example, if the variance of n'th DCT coeff (taking its values from
> different blocks of 8x8 size) is greater than the same of m'th coeff,
> it
> should come before it.
> abhishek.
> f...@hotmail.com wrote:
>
>> Hi guys,
>>
>> I am currently doing a project about the multiple description coding
>> for a image using correlating transform. In one of the IEEE
>> journals, the author describes the encoder as
>>
>> “The coder is based on a modification of a standard JPEG coder. In
>> our baseline encoder, each block of samples is transformed using an
>> 8*8 discrete cosine transform (DCT). These coefficients are ordered
>> to have decreasing variance.”
>>
>> I have done the DCT part, but can’t figure out how to rearrange
>> these coefficients in the orders of decreasing variance. I thought
>> the variance is only meaningful when it refers to a set of random
>> variables and the value of variance will not be affected by the
>> ordering of variables.
>>
>> Could anyone help me with this problem? Thank you very much.
>>

(You need to be a member of imagedsp -- send a blank email to imagedsp-subscribe@yahoogroups.com )