Grouping of Unconfirmed Face Results

[kartikx]

Hi everyone,

In the past few weeks I've been working on Grouping Unconfirmed Results of Face Recognition. As was suggested by Maik in a bug report, I decided to implement categorization of Faces based on the Suggested Names. This indirectly leads to similar faces being grouped together, as they would have the same suggested name.

Here's the results : https://imgur.com/a/CKVopzi

The results in the Unconfirmed Tags will get split up based on the name suggested by the Algorithm whereas the Confirmed Tags will get split up into Unconfirmed and Confirmed results for that person.

There's a lot of implementation details that I would like to mention, but I will go over those in a later mail. I have committed my changes to my branch, so feel free to test it over the weekend. Let me know what you think of this.

Kartik

[Gilles Caulier-4]

Hi Kartik,

It's a very interesting feature. Really the next digiKam 7.1.0 or
7.2.0 including your code (and also the new IA algorithms for face
engines from Nghia) will certainly rock...

I'm currently finalizing the 7.0.0 release announcement. Later that, i
will be available to review and test your code.

My best regards

Gilles Caulier

Le jeu. 16 juil. 2020 à 19:57, Kartik Ramesh <[hidden email]> a écrit :

>
> Hi everyone,
>
> In the past few weeks I've been working on Grouping Unconfirmed Results of Face Recognition. As was suggested by Maik in a bug report, I decided to implement categorization of Faces based on the Suggested Names. This indirectly leads to similar faces being grouped together, as they would have the same suggested name.
>
> Here's the results : https://imgur.com/a/CKVopzi
>
> The results in the Unconfirmed Tags will get split up based on the name suggested by the Algorithm whereas the Confirmed Tags will get split up into Unconfirmed and Confirmed results for that person.
>
> There's a lot of implementation details that I would like to mention, but I will go over those in a later mail. I have committed my changes to my branch, so feel free to test it over the weekend. Let me know what you think of this.
>
> Kartik

[kartikx]

I'm glad you like it :D, please take your time to review your code, I'm not in a hurry.

Kartik

On Fri, Jul 17, 2020 at 9:42 PM Gilles Caulier <[hidden email]> wrote:

Hi Kartik,

It's a very interesting feature. Really the next digiKam 7.1.0 or
7.2.0 including your code (and also the new IA algorithms for face
engines from Nghia) will certainly rock...

I'm currently finalizing the 7.0.0 release announcement. Later that, i
will be available to review and test your code.

My best regards

Gilles Caulier

Le jeu. 16 juil. 2020 à 19:57, Kartik Ramesh <[hidden email]> a écrit :
>
> Hi everyone,
>
> In the past few weeks I've been working on Grouping Unconfirmed Results of Face Recognition. As was suggested by Maik in a bug report, I decided to implement categorization of Faces based on the Suggested Names. This indirectly leads to similar faces being grouped together, as they would have the same suggested name.
>
> Here's the results : https://imgur.com/a/CKVopzi
>
> The results in the Unconfirmed Tags will get split up based on the name suggested by the Algorithm whereas the Confirmed Tags will get split up into Unconfirmed and Confirmed results for that person.
>
> There's a lot of implementation details that I would like to mention, but I will go over those in a later mail. I have committed my changes to my branch, so feel free to test it over the weekend. Let me know what you think of this.
>
> Kartik