{"id":2159,"date":"2023-10-20T17:44:02","date_gmt":"2023-10-20T15:44:02","guid":{"rendered":"https:\/\/tristanromain.com\/?p=2159"},"modified":"2023-10-21T10:38:11","modified_gmt":"2023-10-21T08:38:11","slug":"dont-be-a-loser-use-qr-codes-with-ai-photo-culling-software","status":"publish","type":"post","link":"https:\/\/tristanromain.com\/index.php\/2023\/10\/20\/dont-be-a-loser-use-qr-codes-with-ai-photo-culling-software\/","title":{"rendered":"Don’t be a loser, use QR Codes with AI Photo culling software !"},"content":{"rendered":"\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t

What are we talking about ?<\/b><\/span><\/p><\/h2>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t

\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

Photographers who shoot high-volume like headshots for\ncorporation, schools or retail fashion are now using more and more integrated\nsoftware based on QR code to speed up their workflow. The QR code is segmenting\nthe shoot by subject, each being uploaded eventually on a dedicated space only\naccessible to the customer interested by the subject. It can be an image of the\nQR code shot which is synonymous with \u201cthe shoot of the subject ends here\u201d or a\ndevice on the camera modifying the metadata adding a QR code to each image, it\ndoes not change the result: a segmentation of the shoot by subject thanks to a\nQR code.<\/o:p><\/p>

There is still a catch though: the shoot is well organized\nby subject, fair enough, but the culling for each subject is still needed. When\na photographer shoots typically a couple of dozens of images by subject, only a\nfew to say a dozen are useful for the clients who tend to prefer having some\nchoice but no need to browse dozens of similar images, and of course no images\nthat are not of the quality needed in terms of sharpness, exposure, or the\ncomposition of the image.<\/o:p><\/p>

\n\n\n\n<\/p>

AI can nowadays help the photographer during this culling\nprocess. But how can AI be helpful for a shoot of hundreds of images segmented\nby subject thanks to QR code ?<\/o:p><\/p>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t

\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t

Why is it so much a problem?<\/b><\/span><\/p><\/h2>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t

\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

High-volume is a productivity challenge for the photographer. \u00a0The photo culling done by an AI software must respect the segmentation by subject, organized thanks to the QR code.<\/p>

\u00a0<\/p>

But it must also be customized to the needs of the photographer \u2013 what is sharp enough for one photographer won\u2019t be to another one, and some photographers want to keep only 50% of the images shot, other only 20%, and so on. Whereas the thresholding of such rules is frequently implemented on AI photo culling software, the segmentation by QR code is rarely done out-of-the box.<\/p>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t

\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t

How to fix it ?<\/h2>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\t\t\t\t

It is a workflow challenge. What AI can do well for photo culling must respect the segmentation by subject. It is not a temporality question, nor a similarity issue (some faces \u2013 like sisters\/brothers can look the same by the way), it is purely a subject segmentation, driven by the QR codes. And the way the software is culling the images must respect this.<\/p>

What are the main features required:<\/p>