Rebuild Faces recognition database without losing Face associations

Does anybody know a safe way to rebuild a Faces recognition database without:
1) disturbing already associated faces / pictures
2) finding new faces
I've already done a huge amount of manual editing and auto-recognition on my 40K photo database and have culled many of the "don't care" faces and hand-identified literally thousands of faces taht reconitions didn't identify. Unfortunately the recognition database appears to be corrupted, manifiesting itself in two ways:
1) Count of suggested recognized faces doesn't match the number iPhoto claims to be there, plus several of the "suggestions" shown have already been previously Confirmed.
2) Crashes that seem to be related to memory leaks - iPhoto use of memory piles up until it rolls over 1.7G on a machine with 4GB RAM, and now over 3.8GB on a machine with 8GB RAM, then crashes.
I'm hoping there is a way to rebuild recognition database based on tens of thousands meticulously identified, and seemingly accurate (no corruption), face associations. Thoughts ?

Thanks Terence,
Was hoping for some miracle, but your thoughts make sense. I'm actually pursuing a 3-fold recovery strategy and will document what works best.
1) Keep going with the current database despite issues. So far, recognition database corruption seems limited to particular faces that have huge numbers of the same person (3-7K), not smaller lists. And I don't appear to be getting cross corruption where I label a person, then they come back with a different name. Only the suggestion lists appear to have problems, and again, only for a few select people. I also upgraded to 8G of memory (vs. 4G) which seems to give me more running room to avoid/postpone crashes from memory leaks. iPhoto databases is still incredibly valuable given all the hand work. When it comes time to add faces I can still get recognition to work more quickly and without weirdness via two methods:
* Start a new Faces entry, with a slightly different name, for a person in one of the "oversized/corrupted" Face entries. Gather up enough faces manually to support reconition, then run recognition for the new entry. Once you have collected/gleened as many faces as one can, merge the entry into the existing, ovesized Face entry.
* and bring any new photos/events in via iPhoto Manager from a clean iPhoto database (see next). Faces come over from the clean database.
2) Develop a new recognition database. I have two physically different iPhoto databases (both Time Machined) that I keep on separate drives. I keep them synced via iPhoto Manager. I recently removed all the "faces" database files from the "shadow iPhoto database" and have begun rebuilding. I now use that as my master for adding/importing new photos and runnign Faces (though Photostream makes that a little trickier). When new events are the way I like, I copy over to the old "corrupt" database using iPhoto manager. Working nicely, but I'm now begining to realize the scope of the naming I did in the "corrupt" database - hundreds of hours to get back to where I am with corrupt database.
3) I'm also trying Picasa recognition on an "in-place" iPhoto database. Of course, there is no way to transfer back to iPhoto, but a fun experiment in recognition nonetheless. Picasa's recognition is similar in concept but works substantially differently than iPhoto. A few things I've noticed:
- Picasa does a kind of clustering of very simialr faces that makes it very fast in some recognition steps. instead of labeling one photo, you label clusters.
- Picasa operates regularly on the whole list of unknown photos, instead of focusing in on recognizing just one person at a time, so one can very quickly wade through confirmation of many people.
- Picasa seems less sensitive to one or two "wrong" persons in a long list of recognized faces vs. iPhoto. Even one wrong photo amonst hundreds appears to steer the iPhoto reconition to pick up many of the wrong person for the suggestion list, while Picasa appears more immune (I would love a utility to weed out a few offending, wrong faces in an already recognized database).
- At the same time, iPhoto appears to go to greater depth in finding all images of a select person when it concentrates on that person.
- Ultimately, both recognizers seem to suffer from "fatigue" and randomness - they get to a poitn where they can recognize now more, until some new random event comes along that frees up or catalyzes additional recognition.
More later

Similar Messages

Redo entire database without losing iphone sync

My itunes is a mess through years of sloppiness. I've got music scattered everywhere, different hard drives, etc... and my database is a mess of duplicates and "can't locate file."
So I deleted the itunes XML database file last night and re-indexed my Mac. (I DID make a backup of the DB first though). Low and behold once I plugged in my iphone to this new, clean itunes it claimed the device wasn't synced to my Mac and I would have to wipe everything off. So I restored the old database....
So without spending money to buy a gimimcky "clean your itunes" programs, is there a way to achieve a fresh database without losing sync?
thanks for any input!

Yes. iTunes keeps separate backups for each device.

Face recognition is not recognizing faces

i just rebuilt my iphoto library today with a mac tech and now my face recognition isn't recognizing faces. any idea what's up?

I feel duped, I bought this new I-Life version specifically for the Faces feature and after giving it 5 good matches for my daughter it spent all night and came up matching virtually every face male or female,even some sculpture and a watch face along with leopards and a few other african animals. Come on, this is pathetic!!!!!!! Can I get a refund?
I had five other family members identified and it listed them as matches for my daughter too....I would have to give up my computer for the rest of the year for it to get a few matches for each.

Rebuild Library from Scratch without losing what I bought

How can I rebuild my library from scratch without losing what I have already bought? I recently upgraded my system drive to a 1TB drive and all the drive letters changed. I had moved my library off of the system drive and I do not remember what the drive letter was. I have all the files just want to rebuild and resysnc. I have XP Pro and a lot of files prior to Itunes opening up to all mp3 devices.

Welcome to AD!
If you have an older iTunesLibrary.XML file, you can open it using WordPad. It might refresh your memory on what drive letter was being used.

Face Recognition for HP Client Security

Hi,
My friend have a laptop with HP Client Security installed. Recently, he download newer version (and may it's lastest ver) but when he install HP Client Security, he must remove Face Recognition first. Because Face Recognition not work with HP Client Security ver 8.3.3.1786.
So we have a newer version of Face Recognition for HP Client Security 8.3.3 ?

Hi,
My friend have a laptop with HP Client Security installed. Recently, he download newer version (and may it's lastest ver) but when he install HP Client Security, he must remove Face Recognition first. Because Face Recognition not work with HP Client Security ver 8.3.3.1786.
So we have a newer version of Face Recognition for HP Client Security 8.3.3 ?

Is there workaround for Face Recognition Failure as # of images grows??

The face recognition features that identified faces to existing tags stopped working after
I had created tagged about 300 images within a catalog containing thousands of images. In searching the internet, I have discovered that other have had similar problems.
When I began using the face recognition feature, when I clicked "Find more items with ..." the program return additional pictures in
the "Unconfirmed" tab. It no longer does so, despite the being numerous appropriate matches.
Also, the auto-fill feature no longer functions. (This referenced the names of the faced identifed) I am also never getting teasers suggesting names, only "who is this."
http://forums.adobe.com/message/2411842 appears to describe the same problems I have been experiencing.

I've just totally given up on face recognition in PSE8. After the recognition system collapsed for one person, it began to collapse for other people who appeared hundreds of times in my photos, and it just became easier and quicker to make "people" tags for them myself, bypassing the face recognition.
In each case the failure was sudden and total, not preceded by an gradual increase in false identifications. One minute it was recognizing my 5-year old granddaughter about 80% of the time, and then suddenly it never recognized her again. It didn't propose her name among the alternate choices offered; instead it proposed 50-year-old men and other totally unlikely candidates. After failing with her, it also began failing with other people who appeared frequently in the photos.
By the way, Picasa 3 went through all the same photos, recognizing them with a high degree of accuracy to the end.
The other annoying problems were that PSE8 kept asking me to identify people I had already told it to ignore, and that it often asked me to identify things that didn't in the slightest even resemble faces, especially hands, but also things such as clouds or even a gingham tablecloth.
I also found that if I had several similar group photos, it would ask me to identify all the people in the group on one of the photos, but then fail to apply tags to the other group photos of the same people. If I exited PSE8 and reentered, it would then proceed to identify faces in the remaining photos. This happened multiple times, and I always checked to make sure I wasn't selecting a subset of the photos.
The face recognition failure is just one of the many problems that have me frustrated with the Organizer. This thing needs a lot of work and it's a big resource hog. In fact, I'm thinking of abandoning the Organizer altogether and just using the editor. It's really a shame. I'm sure that the Organizer is the only interface a lot of PSE8 users know, and it shouldn't have been allowed to go on the market in this state.

Face recognition to log in WITHOUT typed password option?

I'm wondering if there's a way to completely disable type-written passwords for accounts in Windows 7. I want to have everyone who uses the machine log in solely through facial recognition, and not even have the OPTION of ever physically typing a password, like hiding accounts on the login screen or something. Is this possible? Thanks!

Hello.
>>>'m wondering if there's a way to completely disable type-written passwords for accounts in Windows 7.
I think so but I have not tried it as I am the only user on this Toshiba. Get the others to sit in front of the screen and do the "Face Recognition" skit. Then save the results named after the subject.
Then set up the Face mode to start first, I forget where this is but I defeated mine at one time. It ought to be easy to find. I defeated mine as the light has to be proper for a face to be recognized, I move around too much and at times, I am in a low-light room and the software does not recognize me.
I can not lead you now as I am not on the Toshiba. Be good. Jean.

IPhoto 9.5.1 face recognition slow with large number of photos ( 20,000)

I have a new fast Imac with new Iphoto 9.5.1. I have >20,000 photos. When I use face recognition with people who have >50 photos associated with them, the rolling ball continues for 20-40 seconds. I didn't have this with the same number of photos with my old Imac2009 with Iphoto 9.4.3, 2011.
Thanks.

That is probably still the initial setup for face recognition for you new iPhoto version. Once the "Faces" database has been rebuild, the performance should get better. Give it a day or two. If the problem persists, back up your iPhoto library and rebuild it.
Hold down the key combination alt/option- command ⌥⌘ firmly and double click the iPhoto library to launch the First Aid Tools. Keep holding down the keys, until you are seeing the First Aid Panel. Select the entry "Rebuild Database" from the panel and click "Rebuild".

Duplicate Face recognition Boxes Possible Bug????

I am having a heck of a time with face recognition and assigning a name to a face and then putting another box around the same face, telling me its unnamed. The boxes vary the slightest bit in size, I cant "x" out of the unnamed box, so I have to close the named box, then use the present unnamed box and re-type the name of the person into it. This is extremely time consuming and if the name is already present, I am not sure why Aperture is putting another unnamed box around the same face which is some cases is exactly the same size and others is differs my .5mm. Is there a way to turn this 2nd unnamed box off or to ideally remove the unnamed boxes from all the photos that have multiple naming boxes? Thanks in advance!
Message was edited by: rhammerman
Message was edited by: rhammerman

Maybe try rebuilding the database by holding down option-command when starting Aperture. I had some issues with duplicates in Faces and rebuilding the library worked wonders.

Performance, stability and usability face recognition

Hi
I have been testing out the face recognition feature in Lightroom CC and I have some questions/comments
My hardware setup:
CPU: i7-5820K
GPU: Geforce GTX 750 Ti, running Benq 3201 UHD display
RAM: 32 GB DDR4
Disk: 1 500 GB SSD for OS and SW, 1 240 GB SSD for Lightroom catalog, 1 6 TB disk with the photos
I run Windows 8.1 with the latest updates.
I have according to Lightroom 120.000 photos in lightroom, mainly RAW, but also quite many JPG. The photos are organized by year, month, camera on disk. A year as typically from 5.000 to 10.000 photos.
Performance
1) It takes days for Lightroom to go through the photos to do face recognition. That is ok, but according to Windows task manager Lightroom seldom use more than 10% CPU, there is little disk access and there is pleny of available RAM. What is the bottleneck?
2) If I try to handle all the photos together the pc starts lagging and everything goes very SLOW, but still no visable stress on the hardware - still about 10% CPU for Lightroom. It can take minutes to put photos to a named person. If I only take one year at the time the respons times are ok.
Stability
1) When I work with all the photos at once Task Manager says "Not Responding" a lot for the Lightroom entry.
2) The screen goes "white foggy" from time to time indicating that the pc/Lightroom has sever problems
3) The screen flickers sometimes when it is stressed out.
4) If I leave it at "All pictures" it is very often inpossible to use it again, I do not get contact with Lightroom, I have to terminate the Lightroom prosess and start again
5) Other applications works allright, but might lagg when Lightroom is very stressed
Usability
1) I thought face recognition would become better and better as it learned from more and more photos. It does not seem that way. It is quite often not helpful at all - you have to do a lot manually
2) It seems it does not use other photos for face recognition than current folder, that means that when I can not use the root folder due to the performance problems I have to go through the same process again and again
3) The algoritems does not seem to take time line into account, when trying to find who the person is. I think that is something that is important to to to make it more useful.
I have been waiting for face recognition in Lightroom for many years and so far I am not very pleased.
Anybody that have had nice expirience with face recognition so far?
Any workarounds?
What do the Lightroom engineers have to say about this?

I agree with all the observations above. I have a 25,000 photo database, and am running a Dell Precision with 16gb ram. After people tagging for a while, the screen starts to flicker back and forth between Lightroom and other open apps, or the desktop. It won't respond until restart or put the computer to sleep & login again.
The face recognition is very poor in comparison to Picasa. The more faces you identify, the more bizarre and random the suggestions for a match become. All races, genders and with very few apparent common features. I would prefer NOT to have thousands of suggested names to sort through. If the face doesn't match within close parameters, just leave it in "unknown" without suggestions.
I like the way Picasa groups similar faces, then asks "Who is this?"
Auto-complete for people names is maddening. If you enter a space (don't names have a space between first name & surname?), it apparently resets. So if I type in John Martin, it wants to auto-complete as John Martin Smith. (John Martin & Martin Smith being two DIFFERENT individuals already identified). I already have 10 men named David in my file. As soon as I hit the space, (assuming no one has his last name as a first name), auto-complete DOESN'T work at all. So my husband's name, (a very common first name) has to be typed in its entirety each time. Picasa doesn't have a problem with spaces...
I have a huge collection of magazine clippings and sewing patterns. I would like a way to exclude certain folders from the face search. I don't even want to see those faces.
I'm stunned by the huge number of non-human things identified as faces. Once in a great while, I can see why... but I never had this problem in Picasa.
I LOVE, LOVE, LOVE Lightroom, but don't think this feature is up to snuff, at all, no way, no how.

Satellite C660 - Cannot start Face Recognition software

When I try and run the face recognition software I get a window saying to contact my administrator - I have administration status on the the Satellite C660 laptop.
Forum searches I have found have not helped in that any time I try and follow advice I still get the same message. The laptop is new.

Hi Sorry but I seem to have mixed up my username and password and had to create a new one LY2! This has taken longer to reply than I had hoped.
I have not tried either of the two suggestions so far but wonder, from your comments, if the problem is that I have used Windows Easy Transfer to copy information and profiles from my old laptop. I therefore am not using factory default settings, nor am I sure how to do so without losing the transferred files. My old laptop, and desktop from which I also copied files, were Windows XP. Because of the File Transfer I do log in using windows log on a password

LRCC Face recognition - best practices?

Ok so we are all new to the wonderful world of face recognition in LR. I'm trying to work out what would be the best practices for using this.
A little bit of background - I have a catalog of over 200,000 images. In addition to portrait and wedding clients, a significant part of my work is with models and another significant part of my work is theatre photography. I have be wanting some sort of face recognition to help with both for some time.
What are your namining conventions for people? - here's mine:
Ideally I would label people as "surname, firstname" so that I can keep members of a family together in "named people" display, but commas are not allowed in names. Also the professional name of many models doesn't fit that pattern eg "Strawberry Venom" or "Cute as Sin" are to models I have worked with.
I am trying to come up with a sensible naming convention at the moment it is "Surname/ Firstname" for clients, theatre folk and friends/family. Models are still a problem, at present I am thinking of "Surname/ Firstname (model name(s))" While I may not be able to remember the real names of models, I do usually know the names from model releases. This naming will still permit me to filter/find them in the keyword List panel by just entering the model name.
On final addition I am making to this this naming convention is the use of a hashtag suffix to the name: #F for friends and family, #C for clients, #T for theatre/actors and #M for models. This enables me to filter on just models, or just actors, or just friends and familiy. Where people fall into multiple categories I add multiple hashtags. So photos of me would be keyworded with "Butterfield/ Ian #F #T"
Unknown / unidentified people.
What I am not yet certain about is how to handle unknown / unidentified people. Unidentified people fall into a number of different categories.
People I don't know and I am never likely to know (Eg random strangers on the street, local tour guides on holiday, random people in the background etc)
This group is relatively easy to deal with - that is to simply delete the face recognition, End of story.
People I don't know the names of yet but I am likely to find out (Eg actors in a production for which I don't have a programme)
For these people I am making up a unique name using the format "date/ Context-Gendernn" Eg an unknown male actor at Stockport Garrick Theatre would be named as "20150313/ SGT-M01" Although this may appear a complex solution it has a number of advantages. If/when I do learn the name of the individual (Eg I photograph them in a different production) it is simply a case of renaming the people keyword. Creating a unique name and not simple assigning all unknowns to a bucket name will help the face recognition algorithms find this person without it being confused by have different faces assigned to the same name. I am also using the hashtage #U to make it easier to filter the unknown faces when I need to.
People I don't know the names of and there is only a slim possibility of meeting/photographing again (Eg guests as a client weeding)
It feels as though I out to just delete the face recognition and have done with it, and this is what I would do except for thing. Other than manually drawing face regions I have not yet found a way to get lightroom to rescan a folder for faces if you have previously deleted the face recognition. This means that deleting face regions from a large number of people is something that cannot be easily reversed. I might just leave these people in the "Unnamed People" category... at lease until such time as there is a way to rescan a folder or colectoin.
Summary
My practices are still evolving. But I hope these thoughts and idea will help others think through the issues and come up with solutions that work for their situation. I am interested in hearing how other people are using the face recognition system. Especially if anyone is aware of any 'best practices' that Adobe or anyone else has recommended.

Glad it helped.
Yes and no. You can still put the people keywords into hierarchies within the keyword list - you can arrange them just like any other keywords.so you just create a "smith family" keyword and store "john smith" under it. What you can't do is apply BOTH smith family and john smith the the same face.
My use of the hash tags came about because I initially had a top level keyword for models, one for clients, one for theatre peple and one for family and firends. Then discovered that some of the theatre folk were also clients (headshots) and what to do when a friend is also a client. So the hash tag system means a person can be both a friend, a model, an actor as well as being a client! (#T #C #M #F).

Advanced Sort, Filter (color, image size), DUPLICATES, face recognition, GUI customizeability

Hi,
here are some of my Lightroom Feature Requests (maybe later more):
- Filter based on image size. Example: show me all photos with a resolution of 1600x1200 or less.
- More complex filtering. Example: show me all photos from the Collection A that are tagged with keyword 'k1' and 'k2' but not with keyword 'k3' and that are in the date range d1-d2 and were made with camera model X.
- Automatic portrait/human face recognition
- Sort and filter by color! Example: show me all photos that are mostly green (blue, yellow, red, ...)
- Intelligent recognition and optional deletion of duplicates. Calculate some kind of fingerprint for each photo (file size won't do I guess) so that it will be possible to recognize duplicates EVEN if they are of different image resolution or orientation (landscape, portrait), i.e. really based on CONTENT of the image!
- Make the components in the side panels movable so that the user can arrange his/her own panel with the components he uses most and on that side of the screen where he wants it.
Besides that: I LOVE Lightroom. A GREAT product!! Use it A LOT!
Nice greetings from Germany,
Stefan.

I second that.
"More complex filtering. Example: show me all photos from the Collection A that are tagged with keyword 'k1' and 'k2' but not with keyword 'k3' and that are in the date range d1-d2 and were made with camera model X."
This should be relatively easy to do. Most other DAMs in the market, e.g. ACDSee, do have this feature. It doesn't really becomes a real database until you have this feature.
"Intelligent recognition and optional deletion of duplicates."
Another great feature in ACDSee. Somehow it finds duplicates, even if filenames and dates are different, shows the different photos and gives you an option on what to do.

No face recognition in iPhoto on my iPad 2

Hello, thank you for opening this thread.
Today I uploaded the photo for the first time on to my iPad 2. When I open the photos, there are Photos, Albums, Events, Places, but there is no "Faces."
I searched the discussion board but I could not find an answer how to let iPhoto do the face recognition.
Could someone please help me?
Thank you!!

Faces will only show as an option in the Photos app on the iPad if you've synced photos from your Mac which have 'faces' associated with them (http://support.apple.com/kb/HT4236 : 'iPhoto ’09, you can also add your Faces, Places, and Events to your device.', and the manual 'Events and Faces must first be configured in iPhoto or Aperture on a Mac, then synced to iPad') - i.e. it must be associated with photos prior to syncing them to the iPad.
If you've synced photos with Faces to the iPad and they aren't showing, then try closing the Photos app completely : from the home screen (i.e. not with Photos 'open' on-screen) double-click the home button to bring up the taskbar, then press and hold any of the apps on the taskbar for a couple of seconds or so until they start shaking, then press the '-' in the top left of the Photos app to close it, and touch any part of the screen above the taskbar so as to stop the shaking and close the taskbar.
If that doesn't work then you could try a reset : press and hold both the sleep and home buttons for about 10 to 15 seconds (ignore the red slider), after which the Apple logo should appear - you won't lose any content, it's the iPad equivalent of a reboot.

How to reset the Face Recognition on folder to start the search/detection over again.

I was going along fine (though very slowly) and had 5000 photos identified as my son. Hours later I was down to a few handful of photo and did a select all and dragged then to my son's named photo. After that I can still see all the photos of him, but Lightroom does not look to detect face on any of the pictures. I tried Photos--> People and all is greyed out. The first run through it detected them (until I did the Select all) and all the photos are still there. It is like Lightroom does not think there are people in the picture and there are , or it thinks that it already analyzed them and will keep doing so.
Is there a way to tell Lightroom to re-analyse this folder (5k of photos) to look for people again. Lightroom seems to have forgotten the Face Regions?
All the other proper in the 5K photos are designed correctly.

I agree with Zim. When you start Toshiba Face recognition tool click on option register Face.
On the right side click on Next and start to practice the face recognition option. For the first time using the face recognition, practicing is very useful. If you choose skip the recognition will start without practicing.
Toshiba Face recognition tool takes a digital snapshot of the users face to create a unique password by digital mapping of key features of the users face. Please set also password. Password (inputting through keyboard) is the backup solution when user cannot log-on by face recognition through Web Camera.
It is not complicated. Just follow menu on the screen.

Rebuild Faces recognition database without losing Face associations

Similar Messages

Maybe you are looking for