• Welcome to the Lightroom Queen Forums! We're a friendly bunch, so please feel free to register and join in the conversation. If you're not familiar with forums, you'll find step by step instructions on how to post your first thread under Help at the bottom of the page. You're also welcome to download our free Lightroom Quick Start eBooks and explore our other FAQ resources.
  • Stop struggling with Lightroom! There's no need to spend hours hunting for the answers to your Lightroom Classic questions. All the information you need is in Adobe Lightroom Classic - The Missing FAQ!

    To help you get started, there's a series of easy tutorials to guide you through a simple workflow. As you grow in confidence, the book switches to a conversational FAQ format, so you can quickly find answers to advanced questions. And better still, the eBooks are updated for every release, so it's always up to date.

facial recognition

Status
Not open for further replies.
S

sbv3

New Member
Joined
Mar 19, 2019
Messages
20
Lightroom Experience
Beginner
Lightroom Version
Lightroom Version Number
8.2
Operating System
  1. Windows 10
Hi,

I have found multiple alternative applications to handle the facial recognition process MUCH faster than LR does.
Is there any way to outsource? E.g. by exporting what has already been recognized and having the rest done externally and then re-importing whatever was found incrementally?

Cheers,
sbv
 
Sort by date Sort by votes
It all depends on how the alternative software handles the metadata they add. I think it should be feasible via the use of xmp sidecar files.
The weak link is on how the people identification is implemented by the 3rd party software, and if that is compatible with LR seamlessly.

Sent using Tapatalk
 
Upvote 0 Downvote
sbv3 said:
Hi,

I have found multiple alternative applications to handle the facial recognition process MUCH faster than LR does.
Is there any way to outsource? E.g. by exporting what has already been recognized and having the rest done externally and then re-importing whatever was found incrementally?

Cheers,
sbv
Click to expand...
Which applications? How accurate?

Phil Burton
 
Upvote 0 Downvote
Hi, thinking about using Azure cognitive API.
This would allow to enrich Metadata with what is in the pic, as well as the face characteristics, as well as face IDs.
Don’t know how LR stores this in their sidecars (if at all) - need to check, unless you have insight?
 
Upvote 0 Downvote
There are dedicated fields in the xmp standard for the Lightroom people tag. For each person identified in a photo there are width, height of area, startposition of area, name and other fields.

If you only need a keyword associated globally to the picture, with no reference to where the person is, you will have less problems in LR compatibility.

Sent using Tapatalk
 
Upvote 0 Downvote
Let me mess around manually and see how this could work :)
Thanks for the hint and the encouragement!
 
Upvote 0 Downvote
sbv3 said:
Hi, thinking about using Azure cognitive API.
This would allow to enrich Metadata with what is in the pic, as well as the face characteristics, as well as face IDs.
Don’t know how LR stores this in their sidecars (if at all) - need to check, unless you have insight?
Click to expand...
sbv,

It may be that you are looking at writing code or doing custom integration. You will also have to sort through the complexities of Azure, which is really targeted at companies, not individuals, as far as I can tell. More power to you. Keep us posted!!

I recently started to do facial recognition for my entire catalog, and I'm not happy with the very high rate of mis-identification.

Phil Burton
 
Upvote 0 Downvote
sbv3 said:
Hi, thinking about using Azure cognitive API.
This would allow to enrich Metadata with what is in the pic, as well as the face characteristics, as well as face IDs.
Don’t know how LR stores this in their sidecars (if at all) - need to check, unless you have insight?
Click to expand...
sbv,

It may be that you are looking at writing code or doing custom integration. You will also have to sort through the complexities of Azure, which is really targeted at companies, not individuals, as far as I can tell. More power to you. Keep us posted!!

I recently started to do facial recognition for my entire catalog, and I'm not happy with the very high rate of mis-identification.

Phil Burton
 
Upvote 0 Downvote
PhilBurton said:
I recently started to do facial recognition for my entire catalog, and I'm not happy with the very high rate of mis-identification.
Click to expand...

I did find Lr did get better; but slowly compared to other tools I have used for facial recognition. And it still is not as good as the old Picasa.

Tim
 
Upvote 0 Downvote
tspear said:
I did find Lr did get better; but slowly compared to other tools I have used for facial recognition. And it still is not as good as the old Picasa.

Tim
Click to expand...

There is no use whinging over the discontinued Picasa. To my knowledge, Google has not offered the face recognition technology to anyone else or made it available in another Google offering.
I can speculate that one reason that Picasa facial recognition was so effective is that all of the algorithms were processed on a web based server where presumably there was greater horsepower available for making all of the necessary comparisons.


Sent from my iPad using Tapatalk
 
Upvote 0 Downvote
clee01l said:
There is no use whinging over the discontinued Picasa. To my knowledge, Google has not offered the face recognition technology to anyone else or made it available in another Google offering.
I can speculate that one reason that Picasa facial recognition was so effective is that all of the algorithms were processed on a web based server where presumably there was greater horsepower available for making all of the necessary comparisons.


Sent from my iPad using Tapatalk
Click to expand...
I'm sure you are right, but that also raises a HUGE, HUGE privacy concern. Just saying.

Phil Burton
 
Upvote 0 Downvote
clee01l said:
There is no use whinging over the discontinued Picasa. To my knowledge, Google has not offered the face recognition technology to anyone else or made it available in another Google offering.
I can speculate that one reason that Picasa facial recognition was so effective is that all of the algorithms were processed on a web based server where presumably there was greater horsepower available for making all of the necessary comparisons.


Sent from my iPad using Tapatalk
Click to expand...
Just commenting how the state of the art has not moved forward in this case.

Tim

Sent from my SM-J737T using Tapatalk
 
Upvote 0 Downvote
clee01l said:
To my knowledge, Google has not offered the face recognition technology to anyone else or made it available in another Google offering.
Sent from my iPad using Tapatalk
Click to expand...

Well, Google Photos uses a really powerful facial recognition and much broader AI engine. You could search for "White Dog" and it would return relevant photos without having to enter a single keyword or tag. Or search for a place (Paris) and it would return iconic photos (Tour Eiffel) from your archives, even from files without GPS coordinates.

Sent using Tapatalk
 
Upvote 0 Downvote
Upvote 0 Downvote
Upvote 0 Downvote
tspear said:
I have used a trial version. Very cool tech; and I am interested to see where they take it.

Tim
Click to expand...
Tim,

I also think it's cool, very cool. But, , but, but, quoting from the Excire website. I've marked words in italics for emphasis.

The Excire Search Keyword Transfer allows for transferring Excire keywords to the Lightroom catalog either interactively, image by image, or in batch mode. Excire Search creates its own keyword hierarchy to not interfere with already existing Lightroom keywords. Once the keywords have been transferred, all the Lightroom filters and search functions can be used to work with the Excire keywords. Note that Keyword Transfer is only available in Excire Search Pro.

To me, the fact that Excire's keywords do not integrate with my keyword hierarchy is a major issue, and why I haven't purchased this software already.

Phil Burton
 
Upvote 0 Downvote
Phil,

Why would you want it to integrate?
I liked the separate hierarchy so I could see what I did vs Exxire.

Tim

Sent from my SM-J737T using Tapatalk
 
Upvote 0 Downvote
tspear said:
Phil,

Why would you want it to integrate?
I liked the separate hierarchy so I could see what I did vs Exxire.

Tim

Sent from my SM-J737T using Tapatalk
Click to expand...
Tim,

You are right, because I would want some mechanism to compare my own keywording vs. Excire's keywords. But Excire might use a synonym of one of my keywords to identify the same part of an image, and then I would have "competing" keywords to identify the same feature within the image. That would be very confusing and, to me, unworkable.

If Excire had some way to allowed me to "train" Excire's use of keywords, then I would be a likely customer. Since Excire is based on AI techniques, it should not be that hard for Excire to implement such a feature.

Phil Burton
 
Upvote 0 Downvote
@PhilBurton

Go the other way. Write a smart collection which both keywords; and use it to remove yours. You no longer need yours unless Excire misses one.
You can then also do the same kind of logic to find the ones you tagged but Excire did not. Or eval the ones Excire tagged and you did not.
 
Upvote 0 Downvote
tspear said:
@PhilBurton

Go the other way. Write a smart collection which both keywords; and use it to remove yours. You no longer need yours unless Excire misses one.
You can then also do the same kind of logic to find the ones you tagged but Excire did not. Or eval the ones Excire tagged and you did not.
Click to expand...
Tim,

That's an interesting idea. I should really evaluate Excire and try it out on some photos that I have already keyworded.
 
Upvote 0 Downvote
Status
Not open for further replies.
Back
Top