KEYWORDS for people
Updated 4/24/2022
In LR there are many ways to do the same thing, each with pros and cons. Some make more sense to some people and other methods make more sense to other people. Here is what makes sense to me and what I use.
Note: As of version 6, (and CC 2015), LR offers a “face recognition” feature. Just be aware that when you use the LR Face Recognition tools the end result is a Keyword for each person you care about. Prior to LR6, this was a manual process. The changes brought in with LR6 is that LR itself helps find faces in images and suggests who each person is through face recognition programming. But, however you do it, either with our without using the Face Recognition tools in LR, the net result is that there are keywords that represent people. And that’s what this blog is about.
Note: I strongly recommend that you use full names with dashes or underscores between first and last name because at some point you’re going to have duplicate first names as well as duplicate last names among the people you’ll be adding. By following the example (note check boxes), I can search for “Dan-Hartford” and get only that one person rather than all the other Dan’s and/or Hartford’s including images shot in Hartford CT.
Snap00
The problem
At some point in time you will have so many images of people you know that you start having trouble finding them when you want to. Usually this happens with photos of family members and close friends but can extend to co workers, and famous people you encounter (e.g., My Selfie with Obama).
What is typically desired is a not too complicated way to be able to find photos that have specific people in them and even better specific combinations of people. For example, find all the photos that have BOTH Fred and Betty in them, have ONLY Fred and Betty or any other combination or people.
Concept of the solution
The key to this problem utilizes keywords, text filters, and smart collections and is really quite simple once you get the concept. What we’ll do is create a keyword for each person we want to track by name. These keywords of individual people may be nested into groups such as Family, Co-worker, Politicians, Performers, etc. but all must be under a common parent KW which in my example is KW “ALL-PEOPLE”. In my example we using the “x” family with 4 members, April, Bob, Charlie, and Dawn.
Snap01a
Assigning the Keywords
This is not as difficult as it might seem, even if you have, literally, thousands of images. Perhaps you even have some prior keywords set up (like one keyword for each combination of people) to help you zero in on the photos you want. Or, you can use the LR Face Recognition tool (as of LR6) to help find them or just add them manually. Then, using the Grid view in the Library module, simply click, shift click, or Ctrl/Option click to select images that contain a certain person and assign that persons keyword to those images. Then go on to the 2nd person, etc. In this way, if you have an image with 3 people, it you will get 3 keywords – one for each person.
Set Up
Once your images with people have been assigned keywords (with or without face recognition) you can use filters or smart collections to quickly retrieve images of the people you’re looking for if you set it up properly.
Start with the “ALL-PEOPLE” keyword, then within that create a KW for each grouping of people that interests you. Some examples are a group for your immediate family, one for extended family, one for friends, one for Co-Workers, etc. In other words look at the people you wish to identify by name and see what sort of natural groupings they tend to fall into. I name these groups with “KWList” at the end of the name to assure they are unique. Then place a KW for each person within their group. This whole thing will work a lot better if you use full names with hyphens or underscores instead of spaces. For example “Fred-Smith” rather than “Fred Smith”.
When talking to many people on this topic, it turns out that many want to be able to include or exclude images depending not only on the named people they know but also the existence (or non-existence) of other people in the photos. To accommodate this, you may want to add two more “names” to your structure within ALL-PEOPLE but not in one of the sub groups. These two pseudo people are “Unknown-Person” and “Crowd”. An Unknown-person is someone prominent in the image that you just don’t happen to know the name of (or maybe don’t care). “Crowd” is the existence of a crowd of other people in the image, like at a concert or sporting event.
Snap01b
For complex mix and match selections you’ll also need two more KW’s that are not under ALL-PEOPLE. One for people you want to include and one for people to exclude.
For my examples, I’ll be using this people structure which contains 4 known people (the “x” family) as well as the extra keywords I just discussed.
What follows are examples of various mix and match scenarios to demonstrate the concept.
Case 1 - One Named Person (text filter)
(Regardless of other people in the image)
This is the easiest. Just open your filter bar (“\” speed key toggle the filter bar on or off). Select Text filter type tab then “Keywords -> Contain -> the name of the person”. Here’s an example looking for all images that have April in the image. As soon as you finish typing the persons name those images will show up in your grid
Snap02
In my example images, the first letter of each persons name is on the cartoon figure in the images. “U” is for unnamed persons, and “crowd” indicates that there is also a crowd of people in the image.
For this test case, you could also click the right facing arrow to the right of the name “April-x” in the Keyword list. This will create a metadata filter (rather than a text filter) and will accomplish the same thing but for various reasons I prefer typing in the text filter.
Case 2 - Any of Several named people (text filter)
(Regardless of other people in the image)
The process here is the same, except you use 2 or more of the specific people keywords in the text filter. For the text filter use “Contain” to get images that have ANY of the named people. In the example below, images that have either April or Charlie or both are found.
Snap03
Case 3 - All of several named people (text filter)
(Regardless of other people in the image)
Use “Contain All” operator in the text filter to get images that have ALL the named people. In the example below an image must have both April and Charlie to be shown.
Snap04
As you can see, all the images have both April and Charlie.
Mixing and Matching People
The above examples are simple cases that most people already know how to do. Where the trouble comes is when you want to do more complex mixing and matching some with exclusions. For example I want all images of Dave as long as they don’t also have his ex-wife Darlene in them. Or, I want images of Dave and his new wife Debbie, but just the two of them – no one else can be in the image.
To accomplish these sorts of complex cases we’ll need to establish two Smart Collections as well as making use of those special KW’s I mentioned at the top of this blog and in some cases also a Text filter. As a reminder, a Smart Collection is a group of images that meet the criteria set forth in a set of rules. You create the set of rules then LR automatically keeps the collection populated with the images that meet that set of rules. As discussed above, we need two parent keywords which define lists of people to INCLUDE and to EXCLUDE respectively. We drag the KW’s for individual people to one or the other of these parents and then use a smart collection to show the images.
Here are several different cases of mixing and matching people.
- Case 4 - Images that have anyone on the INCLUDE list with or without others as well
- Case 5 - Images that have anyone on the INCLUDE but without anyone on the EXCLUDE list
- Case 6 - Images that have anyone on the INCLUDE list but no one else
- Case 7 - Images that have all on the INCLUDE list but no one else
Smart Collections for complex selections
PEOPLE: INCLUDED minus EXCLUDED
This Smart Collection will return images where one or more of the people on the “INCLUDE” list are in the image as long as non of the people in the “EXCLUDE” list are in the image
Snap05
PEOPLE: INCLUDED EXCLUSIVLY
This Smart Collection will return images where one or more of the people on the “INCLUDE” list are in the image as long as no one else is in the image as well. “No one else” means no one listed with a keyword that is within the “ALL-PEOPLE” parent keyword
Snap06
Case 4 – Any INCLUDE list person
(Regardless of others in the image)l
We saw this already as case 1 and 2, but let’s do it a different way which in many ways is easier than typing names in a text filter. This time we’ll use our INCLUDE keyword list and one of our new smart collections. All we do is drag the KW’s for the desired people down to the “PEOPLE-INCLUDE” parent and then click on the “PEOPLE: INCLUDED minus EXCLUDED” smart collection. In this case I dragged April and Charlie down to the “INCLUDE” list.
Snap07
The result is all images that contain either April or Charlie, regardless of any others in the image.
Snap08
To add another person to the mix, just drag their KW to the INCLUDE parent. To remove someone from the mix, drag their KW back up to “”OurFamilyKWList” parent
Case 5 – Any INCLUDE list person minus EXCLUDE list people
(regardless of others in the image)
This case is done the same way as Case 4, except that we also drag the KW’s for people we want to exclude to the EXCLUDE list. As we left Case 4, we are seeing all images that have either April or Charlie. But now we want to exclude images that also have Bob in them. So, we just drag the Bob KW down to the exclude list. Now we’re seeing all images that have either April or Charlie in the image as long as Bob is not also in the image. Compare the screen shot below with the Case 4 screen shot and you’ll see that images that had Bob in them are now excluded
Snap09
Case 6 – Any INCLUDE list people but no one else
So let’s take this one step further and get all the images that have either April or Charlie but only if they don’t have anyone else in the image as opposed to just not having Bob. First we’ll drag Bob back up to where he belongs, leaving April and Charlie in the INCLUDE list. But this time we use the “PEOPLE: INCLUDED EXCLUSIVLY” smart collection
Snap10
Now you can see that we have images that have either April or Charlie but only if the image does not have any other people from the ALL-PEOPLE parent. You’ll notice that images with KW “Crowd” are being excluded as “Crowd” is under ALL-PEOPLE. If we decided that April or Charlie as part of a crowd is OK then we just drag KW “Crowd” out from under ALL-PEOPLE (for example into “PEOPLE-EXCLUDED” just to keep “Crowd” nearby to drag back later). Now we have April or Charlie without other folks except for “crowd”.
Snap11
Case 7 – All INCLUDE list people but no one else
This is actually the same as case 6 except we add a text filter for “Keyword -> Contains All -> (names)”. The list of names in the text filter are those that must all be present in the image. So, continuing where we left of in case 6, if we only want the images that have both April & Charlie (with or without the crowds), we add a text filter listing April and Charlie. As you can see below, we now just have the images that contain April and Charlie, with or without the crowd but as long as no one else in the “All-People” hierarchy are also in the photo.
Snap12
Conclusion
As you can see, simply dragging keywords in and out of the INCLUDED and EXCLUDED keyword parents and picking one of the smart collections gives you a broad spectrum of mix and match options for folks which can be further narrowed by use of the text filter.
At this point, I’m hard pressed to come up with a likely case that this model doesn’t handle. Of course there are some off-the-wall cases like “must have April and Charlie but not if both Bob and Dawn are also in the photo” (Bob without Dawn is Ok as is Dawn without Bob), but maybe one of my readers will rise to the challenge and point one out.
When you’re done with your mix and match searches, don’t forget to drag your people keywords back to where they belong.
