While there are certainly alternatives for tagging / AI facial recognition, object identification, etc. my point is that Qnap is misrepresenting their product as useful for these things, when in fact - the effort a user puts into those features, training the AI, populating the tags and people, is entirely at risk. They don't disclose this risk - and lead users into a very costly trap. That's not the fault of the User, that's Qnap's misrepresentation of their features, by commission and omission.Theliel wrote: ↑Tue May 18, 2021 6:50 pm The main problem here is divided into two parts, one part is the fault of the user, another part of QNAP for not posing it in a much more efficient and intelligent way.
-User
Using a NAS's own Software (whatever it is) to sort / tag tens of thousands of photos is nonsense. There is a multitude of infinitely more efficient and faster software for this. QuMagie works generally well, but cannot compete with applications that have been in the market for many years maturing. Above all, and this links with the second problem, when a large amount of Software does not make the changes in the metadata of the images, but in a DB or independent files, as it does for example GPhotos, QuMagie, Synology ...
-QNAP
Nowadays, the metadata of the images allows practically any type of characteristics to be added to them. To give an example, but not exclusively, facial regions (and labeling)
-------------------------
This is essential, because if QuMagie could read (and interpret) the metadata of the images), thanks to the EXIF standards for facial regions, labels and others, not only the changes made by QuMagie would persist between different systems, but any software could read them without any problem, and more importantly !! On the contrary also, one could load his entire collection already tagged with any other software and QuMagie perfectly interpret people and others.
Still, it would always be necessary to do a full scan of the entire library and build the training data for faces and object identification from scratch, but this would be totally unattended and automatic, take longer or shorter.
------------------
Personally, I use DigiKam to organize / catalog / tag. Any change made in the labeling of people or of any kind in general, the EXIF standards are used, so if I open said images in any other software, it identifies it perfectly, and work pretty well with QuMagie, except Facial Tagging
QuMagie PERFECTLY detects and interprets the normal tagging of EXIF photos, but unfortunately it does not support the EXIF tagging of facial regions (are interpreted like normal tags, well, at least its something), for me it is the only limitation it has, the day that QuMagie is able to switch to its facial AI engine the EXIF data of facial regions of the photos already tagged, it will be really the king software
The tagging and facial recognition part was tempting to me because it gave me a way via any device (laptop, phone, tablet, etc) to quickly find / share / post / edit / iterate a photo using the app on any of those devices, or in the web-interface.
I need that integration, particularly after Apple (insanely) removed photo tagging from iOS.
Having that integration with my means of storage, backup, personal cloud access, tagging, etc. - on my own NAS - is exactly why I bought a Qnap device.
I don't want any cloud service provider, tech giant, social media censors, government, between me and my media library, and I want that library quickly searchable - as well as organized by virtue of tagging as the media is added to the library incrementally.
You mentioned DigiKam. I'll check it out, but given what I've described as a use-case, do you think it can work?
I want an end-to-end personal solution, no middle-men, no company oddly taking away the most useful and necessary features to keep a media library organized, reliable, and accessible.