Face Regions Metadata and LR6

[gcoupe]

I see that Lightroom 6 now has support for face recognition and face regions. Even better is the fact that Adobe has used the Metadata Working Group's standard for image regions (just like PSU and Picasa). From the LR support forum:

When you export a photo, unchecking the option Remove Person Info, the names are also written to the IPTC Extension Person Shown and the MWG Regions too.

[Mike Buckley]

Sorry, Geoff, but I don't understand what any of that has to do with using Photo Supreme.

[gcoupe]

Simply because my workflow involves using both PSU and LR, and now I can use the automatic face recognition of LR6 to help automate tagging people, and know that the metadata will be understandable, and usable within PSU...

[weidmic]

Geoff,
Do you know if its possible to assign a label automatically, if a name tag has been found?

Cheers,
Michael

[gcoupe]

Do you mean in PSU or LR6? The latter is possible (after user confirmation) - there's a video tutorial here: https://helpx.adobe.com/lightroom/how-t ... hotos.html

I don't think it's possible in PSU, because there's no automated face recognition, everything is manual.

Bear in mind that I've only just installed LR6 - I'm still finding my way around how I can adapt my workflow to take account of this new feature.

[Hondor]

- I'm still finding my way around how I can adapt my workflow to take account of this new feature.

Any hints of a feasible workflow involving BOTH Lightroom AND Photo Supreme concurrently would be most appreciated. I really like the Developer Module of LR, but I'm a bit stumped over how to use it without maintaining a duplicate catalogue database for LR and PSU respectively. Since the databases with previews more or less takes up space corresponding to a quality-sized JPEG-copy for each photo, I think the combined size of both catalogues is a bit of a stretch for those minuscule SSD-disk, they offer in machines nowadays. I'm not really ready to go down the NAS-route yet.

[Mike Buckley]

I really like the Developer Module of LR, but I'm a bit stumped over how to use it without maintaining a duplicate catalogue database for LR and PSU respectively.

I rarely use Lightroom but when I do I export the edited file as a full-size JPEG and I then remove the original file from the catalog; I am using Lightroom only to edit photos, not to catalog them. If I were using Lightroom as my primary image editor, I'm reasonably confident that I would do the same.

Unlike a lot of people, I have no need to save the edit steps once I have finished editing an image. I custom post-process every image (rather than use batch processing) and I rarely need to revisit an image. The very few times that I do need to re-edit one, I'm happy to start over from scratch. I have done that no more than five times out of my more than 10,000 images.

[Hondor]

custom post-process every image (rather than use batch processing) and I rarely need to revisit an image. The very few times that I do need to re-edit one, I'm happy to start over from scratch.

Well, consider me a niggler - I frequently re-process my abysmal pictures in an effort to make them moderately viewable. Also, I'd really like to archive in (lossy) DNG-format in order to keep my edits in-file and expanding my options beyond JPEG.

[blinkit0]

Simply because my workflow involves using both PSU and LR, and now I can use the automatic face recognition of LR6 to help automate tagging people, and know that the metadata will be understandable, and usable within PSU...

Geoff,
I am also very interested in the face recognition LR feature and how to integrate it with PSU.After using the face recognition in LR I can see the LR tag in PSU but it shows as a new tag even if it already exists in PSU. In other words if I use Juan in LR face recognition PSU can see the Juan tag but will no match with the Juan in PSU.

I also use LR for development only and PSU for ingestion and cataloging. The LR facial recognition is pretty fast and fairly accurate. Would love to find a way to use it with PSU.

I am also interested on seeing a few LR/PSU workflows from PSU users.

Juan

[lippe]

As i shoot raw+jpg, I tried this:
I have Auto sync on in PSU. When i use LR to find Face regions and save the meta from a raw (cr2), PSU picks up the regions with person info and can display the region in the PSU image editor.

[gcoupe]

Up until now, I've always done my metadata handling in PSU, and never created metadata in LR. However, if I am to use the automatic face recognition feature of LR6, then I would need to let LR6 save metadata in my images.

Unfortunately, I can't seem to persuade LR6 from messing about with my keywords.

I use keyword hierarchies, and in PSU have now switched to reading and writing LR hierarchy metadata. That means that the end node of a keyword hierarchy gets written to dc:subject, while the complete hierarchy is written out to XMP using the lr:HierarchicalSubject.

So, for example, a photo tagged with the keyword "Christmas" will have a dc:subject entry of Christmas, while the LR entry will have Events|Holidays|Christmas.

All, fine, and LR6 will import the photo into its catalog quite happily, and show the keyword "Christmas", and also know that this sits in a keyword hierarchy of Events|Holidays|Christmas.

However, if I then tell LR6 to save metadata to the photo, it insists on writing out all the levels of the hierarchy as separate keywords into dc:subject, as well as preserving the lr:HierarchicalSubject entry. I don't see any way to turn off this behaviour, and unless I can do so, I cannot use LR6 to write out metadata...

[DirkS]

However, if I then tell LR6 to save metadata to the photo, it insists on writing out all the levels of the hierarchy as separate keywords into dc:subject, as well as preserving the lr:HierarchicalSubject entry. I don't see any way to turn off this behaviour, and unless I can do so, I cannot use LR6 to write out metadata...

For me it's the single most annoying 'feature' of LR and AFAIK it's not possible to switch it off.
I *do* want to write metadata & previews back to my DNGs.

I have switched off reading / conversion of keywords in PSu (permanently).
I let LR write the metadata / previews and after PSu has read the info I let PSu write the metadata again to restore the keywords.
Not very user-friendly, but it works for me.

[tstoddard]

For me it's the single most annoying 'feature' of LR and AFAIK it's not possible to switch it off.

There is always at least one problem that arises when trying to use more than one application to manage files. I have often thought about how great it would be if PSU were to work together with one of the major raw processors and integrate the two together. I imaging what it would be like if I could click on the adjust button and have the PhotoNinja sliders show up instead of the PSU's adjustment sliders. I know that PhotoNinja can be used as a plugin for Photoshop so it seems that it shouldn't be too much of a stretch to do something similar with PSU.

It would be just as good if LR were to incorporate PSU's functionality for managing metadata but I know that would never happen.

I realize this is off topic but I just couldn't resist bringing it up.

[gcoupe]

At the risk of taking this even further off topic...

I'm just wondering whether I should just bite the bullet and let both PSU and LR6 fly with all the levels of keyword hierarchies written out as keywords. It has the disadvantage of increasing the danger of not being able to see the wood for the trees (instead of single important keywords, you get everything showing up), but at least LR6 and PSU wouldn't be fighting each other.

The alternative would be to adopt DirkS' approach.