Adobe Lightroom is my image processing and asset management application of choice. As at version 5 however, it doesn’t provide face recognition, which I want. Picasa does a very good job of that, but by default doesn’t save the face rectangles or names to the image file itself, rather it uses a proprietary database and folder-based text files (picasa.ini) to store the data. The face metadata therefore is not portable should the image file be moved (outside of Picasa), uploaded to the web (except to Picasa Web Albums) or emailed to someone else, which undoes all the work done in recording the faces.
Perhaps Lightroom 6 will have face recognition? (Are you listening Adobe? Even a cursory review of various forums reveals a significant user base that wants this feature, although to be sure there are also many who would not find it useful.) If so this might all become redundant. I would be pleased to re-evaluate and modify procedures and applications to finally migrate everything into one schema!
Evaluating Alternatives
This post describes my evaluation of alternatives to get Lightroom to make use of all that hard work done in tagging faces. Because Lightroom doesn’t do faces, the recognised best outcome that could be achieved was that the face names would be imported as keywords, adding them if necessary, and that the rectangle information and association between the face and that area would be preserved. Thus my aim was to find an application that does a good job of face-recognition and tagging and saves the face metadata to the image file itself. Alternatively, because Picasa is known to do a good job, take the information that Picasa stores and insert it into the image using an intermediate application. In either case, test whether the face metadata translates into keywords in Lightroom, and that Lightroom does not subsequently modify or remove the face and keyword metadata. Subsequently, test that the face metadata can be modified if necessary in the original application (and if necessary the intermediate one as well) and any changes can be easily ported back into the Lightroom environment.
Technique
At this point it might be worthwhile to point out that during this phase of exploring and testing it soon became apparent I had made a mistake when first starting to use Picasa face tagging. As is my custom elsewhere (eg phone contacts) I used the format Lastname<comma> Firstname for names. Because Picasa’s search capabilities are quite good (as you would expect from a Google product!), for photos of females that were of a genealogical nature I extended this to store maiden and married names in the format MaidenLastname (MarriedLastname)<comma> Firstname. For some people with multiple marriages this even became something like MaidenLastname (FirstMarriedLastname<comma> NextMarriedLastname)<comma> Firstname. A search for any part of the name would find the appropriate photos – great! Except that when starting to use these as Keywords in Lightroom, the commas became keyword separators so it became a huge mess :-(. I therefore went into Picasa’s People Manager function and changed the format of all my Picasa contacts into the format LASTNAME Firstname, or MAIDENLASTNAME (FIRSTMARRIEDLASTNAME) (NEXTMARRIEDLASTNAME) Firstname in the extreme example above. This still gives me all the search benefits but avoids the keyword comma problem altogether.
The following applications are known to provide face-recognition and keywording capabilities:
- Picasa (http://picasa.google.com/) (version 3.9.137)
- Photoshop Elements (Version 9.0.3)
- Windows Essentials 2012 Photo Gallery (http://windows.microsoft.com/en-au/windows-live/photo-gallery#photogallery=overview) (version 16.4.3508.205)
No doubt there are others I haven’t looked at, but these three are at my disposal and two of the three are (at least for Windows users) ubiquitous and free. Lightroom version used was 5.3.
The following intermediate applications are known to take Picasa’s face metadata and write it into the image file:
- Andreas Vogel’s AVPicFaceXMPTagger (http://www.anvo-it.de/wiki/avpicfacexmptagger:main) (Version 1.7)
- Cris Lovell-Smith’s Picasa Face Embedder http://cris.lovell-smith.com/downloads/picasa-face-embedder) (Version 0.96)
- Jeffrey Friedl’s Picasa Face Import plugin for Lightroom (http://regex.info/blog/lightroom-goodies/picasa-face-import)
- PicFace (https://code.google.com/p/picface/) (Version 0.9)
In each case the intermediate application only operates using the picasa.ini files together with the names database.
So the first step was to tag some faces in each application in a controlled environment and see whether the face names translate into keywords in Lightroom. Actually it seemed it might be useful to test whether the face tags were recognised by each of the other applications, for completeness. For a “controlled environment” I used a netbook with fresh installations of Picasa, Windows Essentials 2012 Photo Gallery, AVPicFacXMPTagger, Picasa Face Embedder and PicFace to create the files with embedded tags, although this wasn’t feasible for Photoshop Elements or Lightroom.
Picasa has an option (under Tools | Options – Store name in Photo [for all future face tags] and also under Tools | Experimental [for selected or all existing photos]), to write face tags to XMP data. Although they both do the same thing, the former option stops writing the face data to the picasa.ini file, so the two are mutually exclusive. The latter option performs a one-off procedure. Obviously this option might obviate the need for an intermediate application, so both options were tested. If using the first (persistent) option, caution needs to be taken because while moving from the default (picasa.ini) to XMP creates the XMP data from the ini file, the reverse is not true. That means that any changes made after selecting the XMP option would be lost (the picasa.ini files still exist after moving to XMP and thus anything done before selecting XMP would still be present).
To document exactly what changes if any were being written to the image files, Phil Harvey’s excellent and well-recognised ExifTool (http://www.sno.phy.queensu.ca/~phil/exiftool/) was used to extract a full metadata inventory both initially and immediately after each stage of processing (before opening the image in another application). The results of these were saved to text files for later evaluation and comparison.
To quarantine the effects of each application, a new folder was created for each of the 7 tests, ie:
- Picasa ini & ACPicFaceXMPFaceTagger;
- Picasa ini & PicasaFaceEmbedder;
- Picasa ini & PicFace;
- Picasa ini & Lightroom Picasa Face Import;
- Picasa XMP (ie “Store name tags in photo”)
- Windows Essentials 2012 Photo Gallery
- Photoshop Elements
and a unique photo was placed into each. Picasa was used to face-tag all photos except the last two in the above list. For the Picasa XMP test that option was turned on; for all the other Picasa tests it was turned off. Photo Gallery and Elements were used to face-tag those photos. Each other photo was then processed individually using the indicated method and then all were imported into Lightroom to see whether the Picasa face tags were translated into Lightroom Keywords. Although not central to the overall investigation, each photo was also opened in Photoshop Elements and Windows Photo Gallery to see whether the keywords were recognised in those applications. The initial results are shown in the table below:
Face-Tag With –> |
Picasa (ini) |
Picasa (ini) |
Picasa (ini) |
Picasa (ini) |
Picasa (XMP) |
WEPG |
PSE |
Process With –> |
AVPicFace XMPTagger | Picasa Face Embedder |
PicFace |
Picasa Face Import |
None |
None |
None |
Lightroom |
Y |
Y |
N |
Y* |
N |
N |
Y |
Elements |
Y* |
Y* |
N |
Y* |
N |
N |
|
Photo Gallery |
Y |
Y |
Y |
Y* |
N |
Y* |
* Face rectangles were not saved in metadata, or not recognised in application
Some interesting observations are:
- PicFace appear to not work at all, except for converting Picasa face tags into Windows Essentials Photo Gallery keywords;
- Windows Essentials Photo Gallery could recognise the keywords created by everything else except Picasa (XMP), but the keywords it writes couldn’t be read by the other two applications;
- Picasa’s XMP format was the worst performer; its keywords weren’t recognised by any of the other applications
- Jeffrey Friedl’s Picasa Face Import addin to Lightroom was easy to use and effective… if only it saved the face rectangle metadata too!
On the basis that the intent was to use Lightroom, PicFace, Picasa’s XMP option and Windows Photo Gallery were excluded from further investigation.
On the basis that it was desired that face rectangles be preserved in the metadata, (meaning that the association between names and rectangular areas of the photo were portable to other applications) Jeffrey Friedl’s Picasa Face Import add-in to Lightroom was also excluded from further investigation.
The ability of AVPicFaceXMPTagger to incorporate hierarchical keywords (ie instead of hundreds of Lightroom keywords at the top level, all the face keywords could be imported as child keywords of the “People” [or some other user-defined] parent keyword) meant that further investigation was concentrated on this application.
Shown below are the AVPicFaceXMPTagger settings I found to give the best results:
Note that Lightroom (version 5 at least) seems to ignore the “XMP-Lightroom Subjects” tags, so there’s not much point in setting them. In fact LR uses the XMP-Subject tags. If you want to use hierarchical keywords (see below), leave the “Picasa face names” checkbox unticked or you will end up with all the Picasa contacts as top-level keywords as well as children of the parent keyword. Note also the separator needs to be a vertical bar “|” character, not the default forward slash “/”. Use the File | Options save menu to save these settings so you don’t have to remember them next time (or refer back to this post!).
Setting up hierarchical keywords in AVPicFaceXMPTagger proved a slight challenge. Although the documentation is quite good, it brushes over this concept; the author obviously knows the subject well enough but doesn’t communicate that knowledge well. I was about to either give up or email the author when I stumbled across the solution. These are the steps required to set up a basic hierarchical keywords list for the program to refer to when creating Lightroom keywords from your Picasa face tags:
- In AVPicFaceXMPTagger, select the Keywords tab. You should see “F [No filename]” in the left pane and a list of all your Picasa contacts in the right pane. Right-click on the [No filename] and select Create element… from the popup menu. Type your parent keyword (I used the word “People” [without quotes]) into the Name box, then click OK.
- In the left pane you should now see “F” and the file path and name you just selected, followed by “K” and the parent keyword you entered in step (1) above.
- Click in the right pane and press <CTRL><A> to select all the entries in the list. Drag the entire selection to the left pane, then press and hold the <SHIFT> key just before dropping the selection onto the parent keyword.
- All your Picasa contacts should now be shown, indented, under the parent keyword, and the right pane should be empty.
- Click Save and save the file in a convenient location.
The end result should look something like this:
Any time you want to import newly-tagged photos into Lightroom in the future, before doing anything else you should click Open and load the previously-saved file, Only those face tags that don’t currently exist in the file will be shown in the right pane, so you then repeat the above process from step (3).
The final step in the evaluation process was to make a modification to the Keywords in Lightroom, then re-open the photo in Picasa and see whether the new keyword was recognised. This was successful. Then, add a new face tag in Picasa, re-process the photo with AVPicFaceXMPTagger and see whether all was still good in Lightroom; this was also successful.
Conclusion
Readers who have got this far will probably have already reached the same conclusion as I did; that until Adobe implements face-recognition and keywording in Lightroom, using Picasa to do the face-tagging and AVPicFaceXMPTagger to non-destructively convert those face tags into keywords embedded in the metadata of the photo provides a workable and fairly straightforward solution to the issue. Not only are the keywords thus created visible in at least some other applications, but also the rectangle data that associates faces with names is preserved and available for those applications that can make use of it. If only all the software companies would agree on a single standard!
From here we will move on to my workflow and digital asset management. But having finally worked through the keywording issue, I now have to do some more research, reading and thinking before I can finalise that. Stay tuned if that interests you!
A reader, Vernon, emailed me with his problem. I’ve re-created his question and my response here, for the benefit of all readers.
Vernon had about 20K photos already imported into Lightroom that he had also face-tagged with Picasa. Using my suggestions from the original post he had processed the images with AVPicFaceXMPTagger, but after re-reading the metadata, Lightroom didn’t display the names as keywords. When looking at a photo’s Properties using Windows Explorer however, he could see the People tags were in fact present. What could be going on here, and how could he fix the issue without re-keywording all 20K photos?
Vernon sent me a copy of an image that had been processed as described. I used ExifTool to extract a full metadata inventory, and a quick review located the problem: the metadata contained the Picasa face name in the XMP RegionPersonDisplayName tag, but not the XMP Subject tag. He had also sent me a screenshot of the AVPicFacXMPTagger window, which confirmed my suspicion that the settings didn’t match my suggestions.
In the AVPicFaceXMPTagger window, the options in the frame captioned “XMP-MP Tags” relate to tags in the Microsoft Photo space. Having the option ticked as shown in the original post results in the Picasa face data being written to the XMP-MP Region structure, which includes the rectangle coordinates and Picasa contact names. Windows Explorer knows about these tags and uses them to display the data found there.
The options in the frame captioned “XMP-Subject Tags” relates to those tags, that are imported into Lightroom as keywords. Vernon didn’t have this option ticked, so no Subject tags were created.
Despite having set up the Keyword hierarchy as per the original post, this had been ignored by AVPicFaceXMPTagger because keywords were not being written; only the Region data was.
There are two possible solutions for Vernon:
1) If he is happy to have all the Person Names appear in Lightroom as top-level keywords, he can use ExifTool to simply copy the existing data in the RegionPersonDisplayName tag into the Subject tag. The keywords can be dragged & dropped in Lightroom to create a hierarchy later, if desired.
The basic ExifTool command to perform this on all files in the current folder and all subdirectories of it is as follows (NB the double-quotes, leading minus sign and full stop at the end are all required):
exiftool "-Subject<RegionPersonDisplayName" -r .
Note this syntax is supported only by ExifTool version 8.83 or later; earlier versions require extra options to de-construct the Region structure. ExifTool by default creates a backup of any file it modifies by appending “_original” to the file name, before it updates it. Users should carefully check the modified files before deleting any backups.
2) If he wants the hierarchical keyword format embedded in the image files easily, he can re-process them with AVPicFaceXMPTagger, after reloading his Picasa contacts, with just the XMP-Subject Tags options set as per my original post.
In either case, he can then re-read the metadata from the images in Lightroom and the face tags should show up as keywords as we want!
To download ExifTool, or for extensive documentation and examples, go to http://www.sno.phy.queensu.ca/~phil/exiftool/.
Ian,
Thank you for the detailed analysis. It has been very helpful.
I’ve had good success using AvPicFaceXmpTagger1.7 working with JPG files out of Picasa, however RAW files have been more frustrating. I’ve tried CR2, RAF, ORF, ARW and RW2. It appears to process and recognize the CR2 and RAF files which are from older Canon and Fuji cameras. The ORF, ARW and RW2 are all from more recent post 2010 cameras. If I try to individually add those files to the processing list, it doesn’t recognize them (setting the filter to all file types). I’ve set the options to create XMP sidecar files for all these formats, but again the more recent cameras never have sidecars generated.
Have you had any experience using this utility with more recent cameras? From what I can tell AvPicFaceXmpTagger seems like an abandoned piece of software which hasn’t been updated in over 4 years. Their support forum hasn’t had a post in nearly 4 years as well. I’ve also downloaded the most recent exiftool and set the program to use it instead of the built-in one. Any suggestions as the utility of using Picasa as a people tagging pre-processor would be enormous. Thank you.
Hi Larry,
I haven’t tried to use AVPicFaceXMPTagger on ex-camera raw files, but my current thinking (which also bears upon other posts on this subject) is to convert them to dng files first, which the program does handle. I’m leaning towards dng as my format of choice thanks to its universality, and I don’t use Nikon’s software to tweak settings etc at all, only Lightroom which of course loves dng. I use Marc Rochkind’s Ingestamatic (http://basepath.com/new/detail-Ingestamatic.php) to do stuff like bulk renaming, importing from the SD card and invoking the Adobe DNG Converter. When I finally get around to posting the next instalment on this thread it’ll have more details about that. At the moment I’m restricted by recent shoulder surgery but hope to get back to full steam soon!
Thanks for your interest & cheers
Ian
Thank you for the advice. Creating DNG files with the free Adobe DNG Converter tool does solve the problem of AvPicFaceXmpTagger not working with newer RAW file formats. To keep a consistent workflow, I’m only going to put DNG and JPG file types into Picasa for tagging which will then get run through AvPicFaceXmpTagger with predictable results. Just a note…AvPicFaceXmpTagger creates XMP-MP regions and depending on your needs this may not be what you want. Exiftool can be used to convert XMP-MP to XMP-MWG as an additional step after going through AvPicFaceXmpTagger.
Although I sometimes use Lightroom for RAW conversion, it isn’t my target database. I’m using IMatch 5. I’ve created versioning relationships so all this metadata can flow from the DNG or JPG “Master” file to all of the associated files. If I want to keep the original RAW files in addition to the DNG, it’s easy to propagate the tagging and hierarchical keywords from the DNG to the RAW (it creates a XMP sidecar).
It would be nice if the author of AvPicFaceXmpTagger would release the source code and let it become an open source project so it can be updated and enhanced.
Hi Larry,
Thanks for your informative update, I’m happy to have helped in a small way. I was very interested to read your comments about XMP-MP to XMP-MWP conversion, and versioning using IMatch – I’ll be looking some more at that! And yes, totally agree about the future of AVPicFaceXMPTagger.
Great article and the only one that really tackles an important workflow that frankly should have been solved a while ago.
Perhaps this question has been answered, but I don’t find the answer to this problem: I’ve tagged all faces in a 20k collection of private photos (all .JPG) in Picasa. Simply because the interface is most convenient. Next to that, I have manually added the name to the “regular” tags to each of the “People Albums” so the file should have double tagging with the names of all the faces in the photo. I have switched the “Store nametags in photo” option on but have also used the experimental option “write face tags to XMP” several times.
Somehow LR has discovered the “first round” of tagging, but now will not import any new tags through either of the methods “Read Metadata from File” or “Sync Metadata”. I’ve tried searching for a way to “Import Keywords” from the Picasa database, but to no avail. I’ve even used the effrey Friedl’s Picasa Face Import plugin, but it only sees a few tags (e.g. a case of 7 from 2300 tags in Picasa) and I’m not getting any of the regular Album tags (e.g. holiday in spain) that I’ve also added in Picasa.
I’m willing to put in the hours for a technically challenging workaround, as it would take me weeks or months I don’t have, to redo the face tagging again in LR. But I’m afraid I haven’t seen one that fits my situation.
Great if there could be someone with practical advise!
Hi PJ,
Thanks for your post! I’m not sure I can help much. Am I right in thinking all your face tagging has been done with the “store name tags in photo” or “write face tags to XMP” options in Picasa on? If so then that’s part of the problem, because nothing else seems to read those tags except Picasa itself. However there should be a way to re-write them into Lightroom-compatible format using ExifTool – although I’ve tinkered around the edges with this excellent and powerful software I’m certainly no expert. I really want to get to know it better though, because it might solve most of these issues for almost everyone if we can get a handle on it and some working examples up.
Regarding LR not seeming to update the metadata after changing it externally, yes, I’ve experienced this difficulty also, and haven’t yet got to the bottom of it. Anyone else out there have any ideas or suggestions?
Cheers
Ian
Hi again PJ,
Having spent some time over the weekend on Vernon’s problem, I now feel more confident in addressing your issue which has some similarities with his. Can you email me a copy of a typical photo that’s not working for you and I’ll see what I can come up with over the next few days? Send me your email address via the the Contact Me page and I’ll reply with mine…
Regards
Ian
Hello Ian!
Do you have some news? I have the same situation as PJ above…
Br,
Jari
Hi Jari,
Thanks for contacting me! At this stage I’m still waiting to hear how PJ is getting on after a further couple of emails. Nevertheless I can say that because there seems to be a need out there which is only partially satisfied by AVPicFaceXMPTagger, I’m thinking of writing my own utility, using ExifTool in the back end, to move face metadata into or out of the various places it can be found in the IPTC & XMP namespaces, but it probably wouldn’t be before the new year, if I decide to go ahead, due to other commitments. Stay tuned to this website for more info in due course 🙂
Hey Ian,
This all look perfect and is exacly what I am looking for excpet I use a mac.
Any way round this using a mac? I couldnt see a mac version of AVPicFaceXMPTagger.
Cheers,
H
Hi Harry,
Sorry, I don’t think AVPicFaceXMPTagger comes in a Mac version, but it could be worth a quick email to the author, via the web url in the post. Good luck!
Hi Ian,
Just wanted to drop a quick note to thank you for your thorough research! I’m currently deciding how best to organise my photo workflow, and facial recognition is something I’d like to do properly, so I’m sure this will be helpful!
Christopher
Thanks Christopher, I still have more to do but I’m pretty sure I’m on the right track 🙂
I need to write up my latest findings soon!
Cheers
Ian
I read too quickly. You already covered the XMP stuff. The problem is that i have everything in XMP now and no easy way to back it out. Argh!
Hi Nick, thanks for commenting! Actually I’ve been thinking about a similar issue over the last week or so, and I think I’ve come up with some ideas for future workflows, although that won’t help your issue :-(. I’m at a similar stage to you in that I’m considering moving towards shooting solely in raw, instead of jpeg or jpeg + raw. I need to spend some time working through my ideas and testing, hope to put up a new post in a week or so, but it has some similarities with what I suggest below…
Addressing your specific issue, maybe you might be able to resolve it using Exiftool (if you don’t have it, use the link in the original post or just Google it), which can read and write both XMP and DNG files? I don’t know if you’ve tried this, but what about you face-tag a raw file in Picasa (assuming your camera’s format is supported – I know Nikon’s .nef is), saving the metadata to an XMP file. My understanding is that when you then use Adobe Camera Raw to convert to the raw file to a .dng, ACR automatically imports the .xmp sidecar file metadata and writes it into the .dng. You could then use Exiftool to inspect the metadata of the .dng file. That should then give you the information you need to process your thousands of files with xmp data and re-write it into the same format as ACR? I’d be really interested to know whether you think this is feasible, and if you try it on some sample files, what settings you use and whether it all works! To be safe, only experiment with copies though, until you’re confident it works (or otherwise)…
Cheers
Ian
I’ve been struggling with this for years. Been using lightroom for about 2 years now but used to use Picasa exclusively. The other major switch was from JPG to shooting in DNG.
A few hiccups because of that:
– AvPicFaceXmpTagger throws an error that says it can’t work with DNG files. So all my DNG files tagged in PIcasa are kinda outta luck.
– Picasa introduced a “Store name in tag” feature a few versions back. You can also under Experimental –> Write Names to Tag. If you have this done, then avPicFaceXMPTagger doesn’t seem to work. I tested this with a file and my name didn’t show up. HOwever, when I deleted my face and re-added it with the “store name in tag” box unchecked; it did work.
This kinda sucks because I have purposely embedded all name tags in XMP in Picasa, but this doens’t seem to show up in LIghtroom at all.
Argh. I have like 40,000 pictures to tag. It’s an overwhelming task to say the least.
Hi,
what happens with the faces image areas defined in Picasa when cropping the images in Lightroom?
Let’s say I have some raws tagged in Picasa. The raws get cropped in Lightroom.
What happens to the resulting jpegs? I would say the jpegs will include wrong image area definitions and I see no way to workaround.
At least when I want to set the faces in the original raw file.
Best regards,
Sobeos
Hi Sobeos,
Thanks for your comment! Yes, the face rectangles become offset by the amount cropped, so they no longer define the correct area. But this happens even when you crop in Picasa after face recognition. As far as I know, no application resets the face coordinates when cropping. What a nice feature that would be!
Cheers
Ian
Cropping in Windows Photo Gallery (latest versions) correctly moves the rectangles to keep them in the right place, rather than them being offset after a crop. That wasn’t the case with some of their first versions with face tagging.
Lightroom preserves the rectangles and face tags created in WPG when editing JPG files even though it doesn’t understand them. However just as I expected if you crop using Lightroom the rectangles are offset by the amount cropped.
Really wish Lightroom had face tagging, I’m evaluating Lightroom just now and most of the features make it far better than other things I’ve used, but the lack of face tagging and ‘faff’ to get tags from other software (great article on best ways to do this by the way) is a sticking point in my decision.
Hi Rosie, and thanks for your input. Good news about the newest version of WLPG, hopefully one day Adobe will get on board. Not everyone’s interested in face tagging but for those of us who are it is important!
“the newest version” is the same one you list here as using in your review as they haven’t updated it since 2012.
I just downloaded Elements to try since MSFT has sadly abandon WPG and 1) holy crap is this thing clunky to use compared to WPG (and even Picasa) (why is the gallery and editor two entirely non-integrated apps with some things like people editing in one but not the other?????) 2) Nothing is auto recognized in Elements photo manager v13 from WPG or Picasa. 3) Did I mention how clunky these Adobe apps are??
Also, despite having the ‘write to file’ on Picasa 3, I do not see it writing any people metadata to the pictures themselves (viewed with GeoSetter w/ExifTool) unlike WPG which does and of course you can even see in the file properties in Explorer.
Hi, and thanks for your comments!
You have highlighted what I have come to believe is one of the most critical issues in this area. “Free” software just doesn’t cut it! WLPG hasn’t been updated since 2012, nor has Picasa since 2013. There’s no incentive for either of the industry heavyweights to advance their software because they don’t make any money out of it. Bugs go unfixed, forum posts unanswered and as for new features… forget them! In the last 2 weeks my Picasa installation has gone to crap for no apparent reason; it re-scanned my disk and I ended up with scores of “new” contacts in the form of “[Contact name] (100%)” that duplicated existing contacts. I now have to spend many hours iterating through all these junk contacts and resetting the faces to their original, correct contact. I found a number of forum posts, going back as far as 2012, identifying the same issue (and also, worryingly, indicating that it might well reoccur). Now that Lightroom supports face recognition and tagging, the sooner I can ditch Picasa altogether the better. I’ve downloaded Lightroom 6 and it looks OK in this respect, though I’ve not yet allowed it to scan my photos to see how good the recognition is. But in any event I’m prepared to put up with an interface that might not be as slick as that of Picasa or WLPG, so long as it adheres to standards and safeguards my photos, data and many hours of work. I no longer have that confidence in Picasa.
On your second point, yes, in fact Picasa does write people metadata to your picture file, in the XMP-MWG group. “MWG” stands for Metadata Working Group, which Microsoft and Adobe are both members of. If there is an industry standard, this is it! Unfortunately it appears Microsoft in typical fashion decided to attempt to create a standard of their own in WLPG, writing face data to the XMP-MP (for “Microsoft Photo”) group. Of course Windows Explorer supports this proprietary non-standard… why should anything else? If you want to view all the metadata in your photos I can recommend EXIFTool (from http://www.sno.phy.queensu.ca/~phil/exiftool/), or the GUI overlay EXIFToolGUI (see http://u88.n24.queensu.ca/exiftool/forum/index.php/topic,2750.0.html). GeoSetter and many other utilities all use EXIFTool as a back-end.
Oh, I should also mention that WPG is much nicer the Picasa when it comes to tagging people as you can set up a outlook account and link Facebook, Google, Yahoo, Skype, etc. so you have all those names to link to. Google only wants to work with google. Adobe seems to think it is 2005.
Personally I find the compulsory link to Google Contacts a real PITA. One of the reasons I do face tagging is for genealogy, identifying people, many of whom predate Google by up to a hundred years. I have a defined naming strategy that doesn’t fit Google’s idea of the same, but I can’t delete Contacts that Google knows about. I have zero interest in linking anything with any social media. If that’s your thing then fine, just goes to show that one size doesn’t fit all! 🙂