Re: GSOC 2018 Proposal: Improvements to N-Dimensional Image Processing - Color Functions

[Ankush Patel]

Edmund, 

No, it did not. However, I’d still like to continue working on the project. 

Ankush Patel

On Apr 29, 2018, at 7:12 AM, edmund ronald <address@hidden> wrote:

Ankush, 

 Did this GSOC project get funded? 

Edmund

ᐧ

On Mon, Mar 26, 2018 at 9:03 PM, edmund ronald <address@hidden> wrote:

Ankush 

yes please, and the setters and getters, ie. it should be easily possible to read in a profile, mod the data tables and the info fields, write it out.

it should also be easy to create default profiles with vanilla values and then mod them and write them out. 

TO ALL IT MAY CONCERN

I think that the ability to use Octave for processing real world images, and for profiling applications, would greatly benefit from the implementation of Matlab's color and ICC functions, and that Ankush's replies on this thread indicate that he has sufficient understanding of these issues to undertake this work. 

Edmund

ᐧ

On Mon, Mar 26, 2018 at 8:48 PM, Ankush Patel <address@hidden> wrote:

Hello,

 

I’m assuming you are referring to these functions - iccfind, iccread, iccroot, iccwrite, isicc- from Matlab? I believe that integrating these functions would be useful as well because they are missing from the color API. I’ll go ahead and write a special section for ICC.

 

Thanks,

Ankush Patel

From: edmund ronald
Sent: Monday, March 26, 2018 2:37 PM
To: Patel, Ankush
Cc: octave-maintainers
Subject: Re: GSOC 2018 Proposal: Improvements to N-Dimensional Image Processing - Color Functions

 

Hi -

 

 If my credentials for giving advice concern anyone, I used to be an ICC individual member, and as a consultant wrote a camera profiling and profile editing suite so I have some idea of color processing.

 

 I have no idea what the status is with ICC and Octave ...

 

 However, as I remember, Matlab has some extremely useful IO functions and constructors  for reading, creating and writing out color profiles, as well as tagging images with same.

 

 I would recommend that these constructors and IO functions be implemented as a priority, image conversion for a couple of cases should then provide a testable template for completing the job later if the project is too short for a complete implementation of the Matlab color API. 

 

Edmund

ᐧ

 

On Mon, Mar 26, 2018 at 5:30 AM, Patel, Ankush <address@hidden> wrote:

Hello,

I have completed and have had a few friends look at my proposal already. If someone could take a look at my proposal, I would greatly appreciate it! I am open to any suggestions to my current proposal.

I have excluded the private application portion from the application.

As far as the scope of my project is concerned, I plan on implementing the missing color functions; however, is this scope too small? Should I consider implementing more functions from the missing functions list?

Are my implementation details lacking? Should I develop specifications and references to each function?

 

Thank you,

Ankush Patel