Participate as a mentor for Google code in

[Jameer Babu]

Hello, 

I want to participate as a mentor for Google Code in this year from digikam. I assure you that I will be honest in mentoring the fellow students and will do my duty accordingly. I request you to give me a chance. 

Looking for a positive reply from you,

[Gilles Caulier-4]

Hi,

What's your computer science skill exactly ?

Best

Gilles Caulier 

2018-09-17 7:00 GMT+02:00 Jameer Babu <[hidden email]>:

Hello, 

I want to participate as a mentor for Google Code in this year from digikam. I assure you that I will be honest in mentoring the fellow students and will do my duty accordingly. I request you to give me a chance. 

Looking for a positive reply from you,

[Jameer Babu]

Hi Gilles,

I mostly do my coding in C++. I learned some algorithms related to image processing and did a project using OpenCV also. Please let me know any further details if required.

[Gilles Caulier-4]

So you want to contribute to mentor a student while the Google code in event. Right ?

Any whish about the subject to give for the student ?

Did you take a look to the bugzilla entries, where a junior job can be identified and assigned ?

Did you take a look to the source code and read the Doxygen API documentation ?

Best

Gilles Caulier

2018-09-17 11:32 GMT+02:00 Jameer Babu <[hidden email]>:

Hi Gilles,

I mostly do my coding in C++. I learned some algorithms related to image processing and did a project using OpenCV also. Please let me know any further details if required.

[Gilles Caulier-4]

If you know opencv, there are some whish from the users about red-eyes correction algorithm.

Typically, the code detect eyes and fix "red" pupil. It must be improved to :

- to be more tolerant about the color reflected by the pupil. Perhaps a settings (actually the algorithm try to do to all the fix automatically).

- Be able to process flash reflection in Dogs pupil (not red).

Look to these reports for details :

https://bugs.kde.org/buglist.cgi?bug_severity=critical&bug_severity=grave&bug_severity=major&bug_severity=crash&bug_severity=normal&bug_severity=minor&bug_severity=wishlist&bug_severity=task&bug_status=UNCONFIRMED&bug_status=CONFIRMED&bug_status=ASSIGNED&bug_status=REOPENED&component=Faces-RedEyes&list_id=1546299&product=digikam&query_format=advanced

If you need help about the code in DK core, let's me hear.

Gilles Caulier

2018-09-17 11:58 GMT+02:00 Gilles Caulier <[hidden email]>:

So you want to contribute to mentor a student while the Google code in event. Right ?

Any whish about the subject to give for the student ?

Did you take a look to the bugzilla entries, where a junior job can be identified and assigned ?

Did you take a look to the source code and read the Doxygen API documentation ?

Best

Gilles Caulier

2018-09-17 11:32 GMT+02:00 Jameer Babu <[hidden email]>:

Hi Gilles,

I mostly do my coding in C++. I learned some algorithms related to image processing and did a project using OpenCV also. Please let me know any further details if required.

[Jameer Babu]

Hi Gilles.

I have seen some bugs in bugzilla and thought that they can be assigned to juniors. I am searching to find some more. I also looked upon the Doxygen API documentation. Regarding the red-eyes removal algorithm, I have found a paper. Once I feel confident, I will do the changes.

--

[Gilles Caulier-4]

I think there is no need to implement a new algorithm for red-eyes. The existing one can be certainly easily modified and adapted for new context.

Code is here :

https://cgit.kde.org/digikam.git/tree/core/libs/dimg/filters/redeye

It use data file for eyes detection here :

https://cgit.kde.org/digikam.git/tree/core/data/facesengine/shape-predictor

Gilles Caulier

2018-09-17 16:21 GMT+02:00 Jameer Babu <[hidden email]>:

Hi Gilles.

I have seen some bugs in bugzilla and thought that they can be assigned to juniors. I am searching to find some more. I also looked upon the Doxygen API documentation. Regarding the red-eyes removal algorithm, I have found a paper. Once I feel confident, I will do the changes.

--

[Jameer Babu]

Hi Gilles,

I will say what I have understood after seeing the code. In redeyecorrectionfilter.cpp everything is fine till the removing the redness by applying the average of blue and green channels. After this we are storing the results. This results are written through .WriteToFilterAction(). But in red eye correctioncontainer.cpp, I feel there is mistake in FormFilterAction(). 

Please let me know your opinion and mistakes in my analysis. 

As our KDE has selected as an organization for Google Code in, I request you to give me a chance to participate as a mentor.

Looking for a postive reply from you,

--

[Gilles Caulier-4]

2018-09-19 8:26 GMT+02:00 Jameer Babu <[hidden email]>:

Hi Gilles,

I will say what I have understood after seeing the code. In redeyecorrectionfilter.cpp everything is fine till the removing the redness by applying the average of blue and green channels. After this we are storing the results. This results are written through .WriteToFilterAction(). But in red eye correctioncontainer.cpp, I feel there is mistake in FormFilterAction(). 

Please let me know your opinion and mistakes in my analysis. 

Well no. The methods listed before are relevant of filter history recorded in image and database. each filter applied to an image use a settings. This one is versionned to be able to list the changes between version.

The code used to apply red eyes identification and fix is mostly located in filterImage() which call correctRedEye() This code run a separated thread to not block the GUI. The code use DImg digiKam image data container, which is converted to OpenCV, processed, and converted back.

Take a care, 16 bits color depth image are supported (RAW files). Image data is always encoded in RGBA color space, as integer values. See DImg API for details. 

As our KDE has selected as an organization for Google Code in, I request you to give me a chance to participate as a mentor.

 

yes sure. Let's go...

Best

Gilles Caulier