Posted by
Herbie on
Aug 06, 2015; 7:46pm
URL: http://imagej.273.s1.nabble.com/How-to-compare-two-images-tp5013922p5013937.html
Xavier,
that's still interesting...
I'm sure you had a look at:
<
https://en.wikipedia.org/wiki/Peak_signal-to-noise_ratio>
Obviously, you should not encounter great problems if you try to compute
the PSNR (according to the given definition) by use of ImageJ.
For the MSE use the "Image Calculator...", the "Math>Square"-, and the
"Measure"-functionality of ImageJ. The remaining operations consist of
number arithmetics only.
You can record the necessary computational steps to obtain an IJ-Macro
or even a Java IJ-PlugIn.
Good luck
Herbie
:::::::::::::::::::::::::::::::::::::::::::
Am 06.08.15 um 20:34 schrieb Xavier Colomé:
> Thank you for your answer.
>
> I think that we can use Peak signal-to-noise ratio (PSNR) to get a PSNR
> value (in dB) for each image. Higher value means best quality.
>
> Has anyone done any test with PSNR?
>
> Thank you and best regards.
>
> 2015-08-06 10:41 GMT+02:00 Herbie <
[hidden email]>:
>
>> Xavier,
>>
>> this is a highly interesting question that boils down to the question what
>> sharpness really is. In some fields sharpness is a subjective entity that,
>> as such can only be determined by psychophysical methods. Of course the
>> main problem with this approach is, that sharpness will depend on the image
>> content.
>>
>> If you think of sharpness as an objective measure then you are to define
>> it mathematically. If you are able to give a mathematical definition of
>> sharpness, it should be possible to help you further.
>>
>> Best
>>
>> Herbie
>>
>> :::::::::::::::::::::::::::::::::::::::::::
>> Am 06.08.15 um 10:30 schrieb Xavier Colomé:
>>
>>> Hello, I'm wondering if is it possible to compare two images and get if
>>> one
>>> has better sharpness than the other.
>>>
>>> Would it be possible?
>>>
>>> Thank you and best regards
>>>
>>> --
>>> ImageJ mailing list:
http://imagej.nih.gov/ij/list.html>>>
>>>
>> --
>> ImageJ mailing list:
http://imagej.nih.gov/ij/list.html>>
>
> --
> ImageJ mailing list:
http://imagej.nih.gov/ij/list.html>
--
ImageJ mailing list:
http://imagej.nih.gov/ij/list.html