[Digikam-devel] Season of KDE

Gilles Caulier caulier.gilles at gmail.com
Thu Oct 30 17:12:35 GMT 2014


Paper is excelent.

Please add some link in current implementation, which is very well documented :

Wavelets NR algorithm to fix image :

https://projects.kde.org/projects/extragear/graphics/digikam/repository/revisions/master/entry/libs/dimg/filters/nr/nrfilter.cpp

NR estimate analyser to compute NR filter settings automatically :

https://projects.kde.org/projects/extragear/graphics/digikam/repository/revisions/master/entry/libs/dimg/filters/nr/nrestimate.cpp

Note : the algorithm is described in this paper :

https://drive.google.com/file/d/0B7yq-xFihT0_UXQ2RHJ4eTdTNFk/view?usp=sharing

NREstimate is also use in Image Qulity Sorter algorithm :

https://projects.kde.org/projects/extragear/graphics/digikam/repository/revisions/master/entry/libs/database/imgqsort/imgqsort.cpp

The goal of this tool is to parse image to detect quality. In this
case, NR is estimated and image will be tagged automatically, using
Pick Label. Note that other quality factor are also tested, as, blur,
exposure, etc...

There are also few test CLI tools, especially this one to report noise
estimation and detection :

https://projects.kde.org/projects/extragear/graphics/digikam/repository/revisions/master/entry/tests/testnrestimate.cpp

https://projects.kde.org/projects/extragear/graphics/digikam/repository/revisions/master/entry/tests/imgqsort/detectnoise/detectnoise.cpp

Best

Gilles Caulier

2014-10-30 9:19 GMT+01:00 Koushik S <skoushik333 at gmail.com>:
> Hi,
>
>     I finished a draft of my proposal. Here's the link -
> https://docs.google.com/document/d/1DTIqDqyIZNmcb0cUuL1xs4_Do9104h5yhej8LxCzm5Y/edit?usp=sharing
>
> Any suggestions/edits are welcome.
>
>
> On Thu, Oct 30, 2014 at 12:41 AM, Koushik S <skoushik333 at gmail.com> wrote:
>>
>> I researched a more on this topic, and as I understand, the threshold
>> needs to be set adaptively. Here's a paper on the topic:
>> http://homepages.cae.wisc.edu/~ece533/project/f03/yangyang.pdf
>>
>> Let me know if I'm in the right direction. Also, does the proposal need to
>> have all the specific implementation details at this point? Because that
>> would take a quite a while, and it's also very hard to know beforehand, in
>> my case.
>>
>> On Wed, Oct 29, 2014 at 9:39 PM, Koushik S <skoushik333 at gmail.com> wrote:
>>>
>>> I checked out the source code, and it seems there is an estimator
>>> function for calculating noise in the image. I don't really know how I could
>>> improve upon the current implementation. Do you have any specific
>>> ideas/solutions on approaching this?
>>>
>>> On Wed, Oct 29, 2014 at 12:11 PM, Koushik S <skoushik333 at gmail.com>
>>> wrote:
>>>>
>>>> Hi,
>>>>
>>>>     That sounds very interesting, and closely related to my field of
>>>> interest. I haven't looked at the code which does the noise reduction, but
>>>> I'll do so right away, and in addition I'll search for any research papers
>>>> on this topic. Any guidance/suggestions on the proposal is welcomed. I'll
>>>> prepare the draft proposal as soon as possible, since the deadline is fast
>>>> approaching.
>>>>
>>>> On Tue, Oct 28, 2014 at 7:25 PM, Gilles Caulier
>>>> <caulier.gilles at gmail.com> wrote:
>>>>>
>>>>> Hi,
>>>>>
>>>>> There is a function to really improve in current implementation :
>>>>> Noise detection on image to tune Wavelets Noise Reduction settings
>>>>> automatically.
>>>>>
>>>>> Currently, this feature work for some kind of noise and shot condition
>>>>> (for ex with high ISO). I have some pictures with noise which are not
>>>>> detected, and reduction cannot be completed automatically. User must
>>>>> play with settings to try noise reduction effect.
>>>>>
>>>>> What do you think about ?
>>>>>
>>>>> Gilles Caulier
>>>>>
>>>>> 2014-10-28 10:30 GMT+01:00 Koushik S <skoushik333 at gmail.com>:
>>>>> > Thanks a lot!
>>>>> >
>>>>> > Note: The deadline is on 31st Oct.
>>>>> >
>>>>> > On Tue, Oct 28, 2014 at 1:43 AM, Gilles Caulier
>>>>> > <caulier.gilles at gmail.com>
>>>>> > wrote:
>>>>> >>
>>>>> >> Hi,
>>>>> >>
>>>>> >> I will take a look tomorow morning if i can propose a subject for
>>>>> >> Season
>>>>> >> of KDE
>>>>> >>
>>>>> >> Best
>>>>> >>
>>>>> >> Gilles Caulier
>>>>> >>
>>>>> >> 2014-10-27 11:11 GMT+01:00 Koushik S <skoushik333 at gmail.com>:
>>>>> >> > Hello,
>>>>> >> >
>>>>> >> >     I stumbled upon this programme from KDE, and wanted to ask if
>>>>> >> > digiKam
>>>>> >> > was participating in it. I started fixing small bugs, and wanted
>>>>> >> > to work
>>>>> >> > on
>>>>> >> > a project, and this seemed like a good motivator and also a good
>>>>> >> > learning
>>>>> >> > experience.  Can anyone guide me on the project that could be
>>>>> >> > drafted
>>>>> >> > for
>>>>> >> > digiKam?
>>>>> >> >
>>>>> >> > --
>>>>> >> > regards,
>>>>> >> > Koushik. S
>>>>> >> >
>>>>> >> > _______________________________________________
>>>>> >> > Digikam-devel mailing list
>>>>> >> > Digikam-devel at kde.org
>>>>> >> > https://mail.kde.org/mailman/listinfo/digikam-devel
>>>>> >> >
>>>>> >> _______________________________________________
>>>>> >> Digikam-devel mailing list
>>>>> >> Digikam-devel at kde.org
>>>>> >> https://mail.kde.org/mailman/listinfo/digikam-devel
>>>>> >
>>>>> >
>>>>> >
>>>>> >
>>>>> > --
>>>>> > regards,
>>>>> > Koushik. S
>>>>> >
>>>>> > _______________________________________________
>>>>> > Digikam-devel mailing list
>>>>> > Digikam-devel at kde.org
>>>>> > https://mail.kde.org/mailman/listinfo/digikam-devel
>>>>> >
>>>>> _______________________________________________
>>>>> Digikam-devel mailing list
>>>>> Digikam-devel at kde.org
>>>>> https://mail.kde.org/mailman/listinfo/digikam-devel
>>>>
>>>>
>>>>
>>>>
>>>> --
>>>> regards,
>>>> Koushik. S
>>>
>>>
>>>
>>>
>>> --
>>> regards,
>>> Koushik. S
>>
>>
>>
>>
>> --
>> regards,
>> Koushik. S
>
>
>
>
> --
> regards,
> Koushik. S
>
> _______________________________________________
> Digikam-devel mailing list
> Digikam-devel at kde.org
> https://mail.kde.org/mailman/listinfo/digikam-devel
>



More information about the Digikam-devel mailing list