<html xmlns:o="urn:schemas-microsoft-com:office:office" xmlns:w="urn:schemas-microsoft-com:office:word" xmlns:m="http://schemas.microsoft.com/office/2004/12/omml" xmlns="http://www.w3.org/TR/REC-html40">
<head>
<meta http-equiv="Content-Type" content="text/html; charset=Windows-1252">
<meta name="Generator" content="Microsoft Word 15 (filtered medium)">
<style><!--
/* Font Definitions */
@font-face
{font-family:"Cambria Math";
panose-1:2 4 5 3 5 4 6 3 2 4;}
@font-face
{font-family:Calibri;
panose-1:2 15 5 2 2 2 4 3 2 4;}
/* Style Definitions */
p.MsoNormal, li.MsoNormal, div.MsoNormal
{margin:0cm;
margin-bottom:.0001pt;
font-size:11.0pt;
font-family:"Calibri",sans-serif;}
a:link, span.MsoHyperlink
{mso-style-priority:99;
color:blue;
text-decoration:underline;}
.MsoChpDefault
{mso-style-type:export-only;}
@page WordSection1
{size:612.0pt 792.0pt;
margin:70.85pt 3.0cm 70.85pt 3.0cm;}
div.WordSection1
{page:WordSection1;}
--></style>
</head>
<body lang="CA" link="blue" vlink="#954F72">
<div class="WordSection1">
<p class="MsoNormal">I would totally love to see more AI tools included in Digikam. Like, being able to search for objects or the content of a Picture without having to manually tag them. E.g. searching photos of beaches, or pictures containing dogs but not
cats.</p>
<p class="MsoNormal"><o:p> </o:p></p>
<p class="MsoNormal">Well, as long as the processing is carried offline (or with my permission). I don’t want my personal pictures uploded on some server without my knowledge...<o:p></o:p></p>
<p class="MsoNormal"><o:p> </o:p></p>
<div style="mso-element:para-border-div;border:none;border-top:solid #E1E1E1 1.0pt;padding:3.0pt 0cm 0cm 0cm">
<p class="MsoNormal" style="border:none;padding:0cm"><b>De: </b><a href="mailto:caulier.gilles@gmail.com">Gilles Caulier</a><br>
<b>Enviat: </b>dimarts, 28 de juliol de 2020 12:05<br>
<b>Per a: </b><a href="mailto:digikam-users@kde.org">digiKam - Home Manage your photographs as a professional with the power of open source</a><br>
<b>Tema: </b>Re: [digiKam-users] a wild idea or wish</p>
</div>
<p class="MsoNormal"><o:p> </o:p></p>
<p class="MsoNormal">Hi all,<br>
<br>
This kind of photo contents analysis exist already with this CLI tool :<br>
<br>
<a href="https://github.com/scheckmedia/photils-cli">https://github.com/scheckmedia/photils-cli</a><br>
<br>
Photils API is very simple to handle and can be used in a dedicated<br>
digiKam plugin. The image content is analyzed locally, a fingerprint<br>
is computed locally by a AI engine, the photo still on computer. AI<br>
parse the fingerprint and return a list of keywords. Et voilà, you<br>
have the tags to validate in client application.<br>
<br>
We plan to include photils as extra digiKam plugin later 7.2.0. Please<br>
be patient...<br>
<br>
Best<br>
<br>
Gilles Caulier<br>
<br>
Le mar. 28 juil. 2020 à 09:15, <neinei@gmx.net> a écrit :<br>
><br>
> might exist a fork already.<br>
> "Photoprism" seems an alternative using some AI stuff for image<br>
> classification, see:<br>
> <a href="https://docs.photoprism.org/developer-guide/metadata/classification/">
https://docs.photoprism.org/developer-guide/metadata/classification/</a> and<br>
> in more general <a href="https://github.com/photoprism/photoprism">https://github.com/photoprism/photoprism</a><br>
> So, somehow object detection seems to be implemented in some open source<br>
> products.<br>
><br>
><br>
> On 7/27/20 10:48 PM, Daniel Bauer wrote:<br>
> > I have no idea if this is possible, if the necessary AI exits<br>
> > opensource, but as there is face-recognition in digikam, I dare to throw<br>
> > in my idea:<br>
> ><br>
> > google, facebook and others can recognize the contents of images. If<br>
> > it's a house, a portrait, a beach, a tree, if male/female, if she has<br>
> > nipples or if cheeks are visible (no joke, both facebook and google can<br>
> > and do - with lots of errors, though), etc.<br>
> ><br>
> > Would it be possible that digikam can analyze and categorize images and<br>
> > make them searchable by content? So that they are searchable by content.<br>
> ><br>
> > That would be fantastic. Imagine!<br>
><o:p></o:p></p>
<p class="MsoNormal"><o:p> </o:p></p>
</div>
</body>
</html>