Median filtr would be the best choice and determine
the auto correlation co-efficients for determining
corrupted pixels.
regards
Bob

--- przemyslawmackowiak
wrote:

> Hi.
>
> How can I detect corrupted pixels? I don't want
> filter an whole image,
> but only coruppted pixels. I consider 'salt &
> pepper' noise
>
> Thanks Start a New Thread

Hi.

How can I detect corrupted pixels? I don't want filter an whole image,
but only coruppted pixels. I consider 'salt & pepper' noise

Thanks Start a New Thread

Hai,
This method will work for impulse noise. Am i correct?
balabharathi

This can be true if the noise is Salt and Pepper noise. But, for other noise like Guassian noise it cannot be true.
> PRaveen Kumar Y
>
>b...@jasmin-infotech.com wrote: Hai,
> By analyzing neighbourhood pixels information, we can differentaite the noise from the edges. Based on the Complexity of analysis, we will get better results.
>
> If i am wrong, anybody can correct it.
>
> thanks,
> balabharathi
>
> Hi,
> >
> > Iam trying to perform diffusion filter to remove the noise. But my way is I need to select only the corrupted pixels based on some thresholding and diffuse only those pixels. But since both the noise and edges are having highest gradient i don't know how to differentaite the noise from the edges.
> >
> > Any suggestions please...
> >
> > Thanks,
> > Indhu
> >
> > Start a New Thread

This can be true if the noise is Salt and Pepper noise. But, for other noise like Guassian noise it cannot be true.
PRaveen Kumar Y

b...@jasmin-infotech.com wrote: Hai,
By analyzing neighbourhood pixels information, we can differentaite the noise from the edges. Based on the Complexity of analysis, we will get better results.

If i am wrong, anybody can correct it.

thanks,
balabharathi

Hi,
>
> Iam trying to perform diffusion filter to remove the noise. But my way is I need to select only the corrupted pixels based on some thresholding and diffuse only those pixels. But since both the noise and edges are having highest gradient i don't know how to differentaite the noise from the edges.
>
> Any suggestions please...
>
> Thanks,
> Indhu
>
> Start a New Thread

Hai,

Follow this & try once, it may help

1.apply a edge filter (Edge image you will get)
2. Substract the edge from original image
3. Then apply your algorithm

Kindly do correct me if it is wrong

Regards
Ayyams

b...@jasmin-infotech.com wrote:
Hai,
By analyzing neighbourhood pixels information, we can differentaite the noise from the edges. Based on the Complexity of analysis, we will get better results.

If i am wrong, anybody can correct it.

thanks,
balabharathi

Hi,
>
> Iam trying to perform diffusion filter to remove the noise. But my way is I need to select only the corrupted pixels based on some thresholding and diffuse only those pixels. But since both the noise and edges are having highest gradient i don't know how to differentaite the noise from the edges.
>
> Any suggestions please...
>
> Thanks,
> Indhu Start a New Thread

Hai,
By analyzing neighbourhood pixels information, we can differentaite the noise from the edges. Based on the Complexity of analysis, we will get better results.

If i am wrong, anybody can correct it.

thanks,
balabharathi

Hi,
>
> Iam trying to perform diffusion filter to remove the noise. But my way is I need to select only the corrupted pixels based on some thresholding and diffuse only those pixels. But since both the noise and edges are having highest gradient i don't know how to differentaite the noise from the edges.
>
> Any suggestions please...
>
> Thanks,
> Indhu Start a New Thread

Hi,

Iam trying to perform diffusion filter to remove the noise. But my way is I need to select only the corrupted pixels based on some thresholding and diffuse only those pixels. But since both the noise and edges are having highest gradient i don't know how to differentaite the noise from the edges.

Any suggestions please...

Thanks,
Indhu Start a New Thread