Using Facial Recognition to Quickly Keyword People
And we're back with another day of Lightroom CC photo editing. Let's look back at what we've done thus far. Right now, we're in the fourth week, but in the first week, we talked about developing a firm foundation in Lightroom. That's where we made sure you knew how to get all your images into Lightroom and the options related to it. We talked about the overall mindset of working in Lightroom, and we talked about features like catalogs, collections, adjustments, and exporting and printing. In week two, we concentrated on organizing your images and adjusting them, and that's where we talked about managing projects, making your images searchable using keywords, retouching, noise reduction, and more. Today, we're gonna cover facial recognition, so we can quickly recognize people in our photographs. And I want to spend as much time in Lightroom as possible, so let's pop right over to Lightroom and get started. Here I have a series of photographs that contain people. And I want to tag each o...
ne of these people, so that in the future, if I ever want to find any of these images, I can do so just by searching for somebody's name. Now usually, in order to find something by searching with text, a keyboard shortcut you could use is command F. That one's pretty straightforward for "find," and if you were to type command F, you can come up here, it says text, any searchable field, and I could type in what I would like. We're gonna be tagging these with keywords. So if I were to go here to keywords and type something in, I could see if it could find any photos. But unfortunately, we haven't done any work ahead of time, so at the moment, it can't find anything. I'll go up to the top of my screen and choose none. So we can't search for any of these images yet. Let's see how long it would take us to get them keyworded. Well if you look at these images, you'll find some of them would be easy to keyword relatively quickly, because the faces within the pictures are relatively large, and so they'd be easy to identify. But then in other photographs, you'll find that the faces are pretty darn small. So you might need to zoom up on the photos to see who it is. And some of these will be group shots, where, like, for instance, this one has a whole bunch of people. We want to be able to identify them all, and if you think about how much time it might take to do that, it could be considerable. But if you have one of the newer versions of Lightroom, then if you go to the bottom of your screen when you're in the library module, you're gonna find an icon that looks like a little face. Now this is within a bar that's called the toolbar, and that toolbar is something that can be hidden. And so, if you don't find that bar when you're at the bottom of the library module, you can either type the letter T, for toolbar, to toggle its visibility, or if you hate keyboard shortcuts, you can also toggle its visibility right up here at the top of the View menu. Once you have the toolbar visible and you're viewing a series of images that contain people, I can click on that icon, and Lightroom will ask me. If I tell it to start finding faces in my entire catalog, it will literally search my entire catalog. If, on the other hand, I choose the bottom option, then it's only gonna search for faces when I click on that face icon. Therefore, I could manually choose which folders get searched because I could navigate to a particular folder, click on the little face icon, and it would only be searching that particular folder. I'm gonna click on the "Start finding faces in entire catalog" thing, and Lightroom kind of gives me a hint that it did something up here. Did you see it had some movement near the upper left of my screen? Well, what it did is, if I go to this area near the upper right, or upper left, sorry, you'll see it says "detecting faces in filmstrip." That means it's not done looking at all these images. It's searching through each one, looking for what it thinks might be a face. And if I were to click up here, I would have the choice here of face detection, and on the right side, I could pause it. Maybe I need my computer to run as fast as possible at this moment in time, or just for some other reason I chose, I don't want it to be looking for faces at the moment, I could click right there. Then it would be paused, and if later on I wanted to continue searching for faces, I could click there a second time to make it start up again. So at any time, you can go up to your identity plate, which is at the upper left of your screen. That's where you can click to pause or start back up your facial recognition. Now after I've clicked on that icon near the bottom of my screen that looked like a face, it started searching this collection of images to see all the faces. And what it's done is, it's isolated each individual face and cropped it so all I'm seeing is the face. So now, even though that one photograph we were looking at earlier, the faces looked really small in the photograph, well now, it's zoomed up on it and cropped it in so we can easily recognize those faces. You'll notice that, underneath each person, it just has a question mark because it has no idea who these people are yet. And at the top of my screen, you will see it has two sections. One is called "Named People," and below that is "Unnamed," and we're currently looking at the unnamed. Watch what happens if I were to click below one of these photographs, and in case you don't recognize this guy, that's me. That's what I looked like maybe a week before I started teaching Photoshop, which was about 20 years ago, or more than 20 years ago. I'm gonna click right below that, where the question mark was, and I'm gonna type in my name. (typing) The moment I press return, it's gonna move this photograph up above, into that section called "Named People." And now, it is doing a few other things while it had time to look through this folder of images. If I look down here, I see the other people, and, let me just tag my wife Karen here. (typing) But you'll notice that on at least one of the photos, there's a number in the corner. And that means it's found more than one picture that it thinks is the same person. Since there's a question mark at the bottom, it doesn't yet know who the person is, but it thinks that there's more than one. If I click on the number, it can expand it to show me both photos that it thinks are the same. Maybe it's 'cause I'm wearing the same hat, I don't know. And I click it again, and it'll collapse it down. Now, I can click again on that and type in my name, and if I start typing my name, if it's a name I've already used in the past, it will try to auto-complete it. It will suggest it. And if I press return or enter, then I've accepted that, okay? And now you can see up here at the top, it'll start giving you a total next to each name. There are three that it thinks are me, there's one that it thinks is my wife. And eventually down here, and I don't know if it'll have enough time while we're doing this demo, you will find that it will replace the question mark with someone's name. It's the name that it thinks should be there. But it's not gonna automatically tag the person because it wants you to confirm it. But if a suggested name appears down there, not only will you have the X mark, which would mean no, that's the wrong name, but there will be a check mark, which means, yes, let's accept it. Now it'll take some time for it to recognize those faces, and we'll see if it has enough time to accomplish that during our demo or not. I'm gonna make the thumbnails smaller. I'm going to the lower right of my screen, right down here, to adjust my thumbnail size. I can also use the plus and minus keys on my keyboard to change the thumbnail size. So now, I don't want to spend all the time it would take for me to type in the names between all these people. I notice that this is a picture of my wife, and so are all these images all the way up to here. So what I'll do is, I'll click on the first image of that sequence, I'll hold down the shift key, and I'll click on the last image of the sequence, and then, I can just drag this on top of my wife in the list above, and let go, and suddenly, you see that it thinks there are 14 pictures of her. I could also type in the name down there, and it would've done the same. Now if I gave this enough time, it would start suggesting names down here at the bottom. Like I would start seeing Karen Willmore below some of these, but for now, let's try to tag some more images and just work through here. These are all pictures of my wife, and you notice it even finds it in really dark photographs. Grab all those, and I'm just gonna drag them up here on top of Karen. I'll grab a few of me. And then, the more I do this, the more, in the future, if I come back in here and click on the people icon, instead of having question marks below these choices, it would start having names appear. It just takes a little bit of time. Drag this on top of me, a few more of my wife. Drag them up. Now, if you're not sure who's in a picture, or maybe it's a little out of focus, or something else, and you want to be able to see the entire photograph, 'cause it might give you a better idea, 'cause it might remind you, based on the location where the photo was taken, or the other contents of the photograph, you can come up to any image and double click on it. If you double click on it, you will see that it will show you the entire photograph, and it'll put a box on top of the face that it's trying to recognize. I could click up there, this is Rich, and I'll do that, and you'll notice, on occasion, it won't recognize a face. In this case, you notice there's a person, but because they're wearing sunglasses and a hat, it looks like it didn't remember them. So, I'm gonna come down here to the bottom of my screen, and you see a little rectangle that's right here. I can click that to toggle the visibility of those rectangles, and I can come up on my screen and just drag around a face that it didn't recognize. And so, that's Rich's dad, so I'm tagging it with that. I'm tagging it with a word just 'cause I want to show you how you can easily change that word later. You might start with "mom" and "dad" or "bride" and "groom," and later on learn those people's names. You can change them, so I'll show you. But if it ever doesn't recognize a face, all you need to do is double-click on the photograph so you can view the entire contents, and then move your mouse on top of the image, click and drag to define a region, and then you can type your name. At the bottom, again, we have the little face icon. If I click it, that just toggles the visibility of that overlay. So if you're used to keyboard shortcuts, you might accidentally do what I just did, which is type the letter G to go back to the grid, and if you go back to the grid, you'll see the normal photographs. You're seeing them uncropped. And remember, you go down here to the bottom of your screen. That's where you're gonna find the face icon, and the face icon sends you into the mode where it recognizes people. It does have a keyboard shortcut, it's the letter O. Like, "Oh, who's that?", you know? You'd think it'd be P or something, but no. So anyway, we can do this with even more people until we get all of them. Look over here, grab these. And you see how I was able to grab more than one at a time. Now what's nice about this, is once you have gone through and done this, this is also Dave, let's put it up there, this is Rich, he's already up there. Once you get this down, then what's nice is, up here, where you see your list of people, if I were to double-click on that, I'll see all the faces that were in that series, that it thought was that person. And if I double-click on any one of those, I can see the entire photograph again. Double-click again, you get back to this list, and so on. Here, there's the word "people." If I click that, it sends me back a level, like sends me back to the list of all the people. The other thing that is cool about this is, if I come in here and do maybe just a few more, and this is the same guy here, and let's see, this is the other one's son. So now I have all the people recognized. But what's really nice is, if you zoom up on one of your pictures, if it's a group shot, it's now gonna have boxes on everybody's heads. So you can have like, a whole wedding party, and their names are all in there. But you'll notice that sometimes it has difficulties finding faces. If you ever have a face that's sideways, it's not gonna recognize it as a face because it can't see all the elements it needs. It needs two eyes, a nose, and a mouth to recognize it as a face. But that doesn't mean I can't come in here, put a box around somebody else, and type in their name, and continue doing that. Remember you can double-click to switch between the full view and the pictures of just the faces. And then over here, you have the word "people." That will send you back another level. So, we've done quite a bit here getting these things tagged with various names. Now, let's get out of people view. Turn that off, and let's see if we can search for those people. I'm gonna type command F for find, and then I'm gonna type in "Ben." And now, notice that it's finding all the pictures of me. Or, I'll type in "Karen." It's finding the pictures of my wife. Or, who else did we get in there, Rich Mindivaris? Find all the pictures of him, so now, suddenly this can become all searchable. But what it's really doing when you are putting in those names is, it's going on the right side of the library module, and that's where you're gonna find your keyword list. And your keyword list is where you're gonna find all those names. So if you find out what somebody's name is, you know, before you had them called "dad" or "son" and you figure out their names, you can double-click on one of these and change it. And when you do, it'll update any photographs that have that name attached to them. And therefore, if it was originally from one shoot, you tagged them all, one person with the word "bride," later on you found their name, you could change it within this list. And you'll find when you double-click on one of these, that there's a setting within this keyword, and we talk about keywords in a different session, but when we did, we didn't mention this check box right here called "person." By turning on that check box called "person," that means it knows it's a person keyword and that it should consider it when it's judging various photographs to figure out who is in the photograph. Now what I like to do, though, is organize all my people keywords, so that my keyword list stays nice and organized. I'm gonna create a keyword for that. I'll just hit the plus sign over here on the left of the keyword list, and I might call this either "people" or "model," or whatever you would use. I usually just call it "model," like who are the models that appeared within my photograph, and I'm gonna use that. I'm gonna not include it on export, 'cause the word "model" itself is not that useful to search from. And then, I'm gonna take all of these name keywords we have, except for that thing called "model," and I'm gonna drag them on top of the word "model" so they're all within that. Therefore, I can easily collapse it down and keep my keyword list relatively clean-looking. Well the problem is, if I do this in the future and I tag additional people, people that I have not tagged up until now, they would appear on the base level of my keyword list. What I'd really like is to have them always appear within this one called "model." And I can do that. If I right-click on the word "model" within my keyword list, there's a choice here called "put new person keywords inside this keyword." and if I choose that, then you'll find the end of the name of that keyword has an asterisk, and that's your indication that now, if I tag additional people, in the future, it's gonna automatically put them within that area called "model." So in my real keyword list, I have a keyword called "who," and then within that is a keyword called "model," and within that there's all the people. And it can automatically put them in there for me, which I find to be really nice. So then the next time I end up importing some images, and I know they have some people in them, I'm gonna click on the name of the folder. I'll go to the bottom of my screen and click on the face icon, let it sit for just a minute where it finds all the faces, and suddenly, it's gonna start suggesting the names, where, if there's people that I've tagged before, like myself or my wife, it will almost certainly figure out the right name, and all I'll have to do underneath the photo is, there'll be a check box to accept that entry. And I'm just confirming that, yes, these are pictures of myself or my wife. And if there's a number next to the image, I'll usually double-click on the number to get it to expand. Actually, you have to click once to get it to expand, just to confirm that all the pictures it found were, yes indeed, a picture of myself or my wife before I end up confirming that that was the right person. So by using people keywords, and by using Lightroom's facial recognition, we can now make it so finding people in our photographs is much faster. But at the same time, I wouldn't rely on it as much as you might think you want to, because on occasion it messes up. Mainly, you have a side profile of somebody's face, or you just can't see their face, but it's blatantly obvious who's in the photograph, you're still gonna have to manually go in and tag those if you want to be able to find them. But it is something that saves a bunch of time. So I do like that. Do we have any general questions about the face tagging? Yes?
You used the term "my real keyword list." Is this a separate keyword list?
I have an empty catalog here. And during a different session, you find that, if you were to see my real catalog, meaning not this limited catalog, I needed to take these images and somehow make them so they didn't already have these names on them. 'Cause if I went to my real catalog, the one I use day to day, all these pictures were already tagged with people. So I needed to create an empty catalog that had some pictures that had never been tagged with names before, and that's what we're looking at right now. If I were to open my real catalog, which I can do, then it will close that particular catalog and open the one that I work on day to day. And if I do, here is my real keyword list. I mean, this is what it looks like when I work in Lightroom day to day. And here you would find "who" and, in there. Here, I actually did remove the one called "models," 'cause I was prepping for this class, trying to get it so that it wouldn't recognize these people, and I ended up putting it in a separate catalog. But usually, there would be one in here called "model," and that's where it would appear.
Okay, so it does integrate with your regular keyword list.
It does integrate with my regular keyword list. I was just working on a simplified catalog because I needed to have it where it didn't already recognize the people, and I was trying to figure out how to do that, simplify it enough where I could act as if it had never been done.
Great, thank you.
Sure. Finally, you can find me online if you'd like on various social media. If you'd like to see some of the pictures I take of my wife doing yoga, that's what some of the images were there, go on Instagram, to theworldismyyogamat. That's what you'd search through to find it. And of course, my main website is the main place to find out what I'm all about. But this has been another day in our Lightroom CC photo editing bootcamp. We have a total of 20 days, over four weeks of time. When you add up all of that, it'll be a considerable amount of knowledge. You're gonna be much better at Lightroom after watching these. Hope to see you for our next episode.