The purpose of digital asset management is to help you manage your asset library easier, find assets faster and distribute them without headaches.
One of the drawbacks or I'd say a daunting reality, of using DAM is it is a rare modern organization that builds its workflows and infrastructure from the ground up with DAM in mind.
What happens more often is that companies graduate to using DAM after they have realized that whatever current standards and practices they're using to manage their photos and videos are too obtuse.
So, teams find out the best DAM for their requirements, sign-up, and start using it. That's good.
But then they realize that they have hundreds, if not thousands of assets that haven't been properly tagged with metadata over the years. That's bad.
AI-generated keywords can certainly help if you're working mostly with products. But if you're on a team that works with people, that's a lot more complicated.
Tagging a photo with a man or a woman doesn't certainly help when you need to know exactly who's depicted in the photo.
But using face recognition will help.
Today, we will talk about how photo face recognition works in Pics.io DAM and how it can be useful for you and your team.
Let's dig in.
A Short Primer on Face Recognition in Pics.io
You might think that describing what is face recognition is a bit redundant, but considering some nuances about how it works in Pics.io, it is actually quite important.
So, when people think about face recognition software, their thoughts are drawn to Google's reverse image search. Let's say you have a photo of a famous person but can't quite remember their name. So you just paste that photo into Google and, if it finds a match, it will show you more photos and give you their name.
Well, if you do the same in Pics.io, it will come back as an unknown person. Why?
That is because Pics.io's facial recognition doesn't identify the person in a photo, even if they're famous. Pics.io scans the face, creating its profile. You'll have to name the person yourself but, once you do, all your photos in Pics.io with the same face will be identified appropriately.
Why is it done in such a way?
Well, the first thing is the privacy concern. Pics.io doesn't see people, per se. It just uses AI algorithms to create an ID for specific face geometry that you then give context to by assigning any given name to the identified person.
So this means that ID by itself is useless outside of your instance of Pics.io. Unless someone gets access to your Pics.io DAM, there's no risk of generated data being misused for fraudulent purposes. Because even we, the developers, can't peer inside.
Using secure protocols such as Single Sign-on (SSO) is a good way to ensure that that doesn't happen.
Constructing a face recognition search in this way is also a more sensible approach if you consider that majority of use cases are for people with low notoriety that you'd need to identify manually either way.
In any case, I'm just a marketing guy, so it'd be hard for me to elaborate on all the technical details of face recognition implementation. If your team has concerns about security or other things, contact us (email@example.com) and we'll answer all questions here or in a call :)
Face Recognition App Use Cases
Now that potentially botched technical explanation is out of the way, let's talk about why would you use face recognition and how it can help you.
We have a lot of users that work with faces on a daily bases - talent agencies, actor guilds, media, modeling studios, etc. - that have a bunch of photos of different peoples. Because that's their bread and butter.
In their day-to-day work, these are things that they often need to do:
- Find a photo of a person (or multiple) in a specific context/locale to use for promos, PR, media coverage, etc.
- Model releases and licensing. Not all models give permanent licenses to use their likenesses for commercial purposes. They need to keep track of any soon-to-expire model releases for removal/potential renegotiation
- A media publication needs to publish a breaking news article talking about a specific individual. They want to spend as little time as possible searching for the person's photo.
These are just a few scenarios but they can perfectly illustrate how using Pics.io's online face recognition in tandem with other features can make your face-hunting pursuits a lot easier.
Finding a photo of a person(s) in a specific context
Let's say you have 100 photos with a varied assortment of faces. You use recognition on all of them.
Once processing is complete, you'll have 100 photos that look like this:
Open up any of the processed photos, click the box around the face that you'd like to identify, and type in their name.
Now, if you were to go and look at all the other photos that have the same face, it will be properly identified as, in this case, Craig Schleifer.
Note: Craig Schleifer is a made-up name created by name generation tool. Any similarities with real people are coincidental.
You can now find all photos with Craig Schleifer just by typing their name in the search bar:
But it just doesn't stop here. Pics.io's advanced search capabilities allow you to combine multiple metadata types to find results that are as precise as you need them to be.
For example, maybe you need only the photos where Craig is smiling. You can use their name with a Keyword smile to do just that.
Or perhaps you want to find photos of Craig in specific locations. If he's a model in your photo agency, for example there might be different scenarios in which he's posing, like a different location.
This is where, custom fields will be particularly useful.
You can create a custom metadata list, with a name Location, and add a dropdown of all frequently used sets that you utilize. So, even if you need something as precise as "Craig and Jane Doe Smiling Together at the Beach", Pics.io will bring up any photos that match that intricate search criterion in an instance.
Now, here we talked about photo/modeling agencies specifically but you can imagine how these photo face recognition capabilities can be expanded to other niches.
What if there's a local politician that you're writing a piece on, and you want to highlight their appearances on different events across the city? The same logic will apply because Pics.io is not a rigid system. It's just a tool in which you get to create your own content and metadata taxonomies to serve your particular needs.
This is also where websites can be helpful as the cherry on top. Let's say you've found all photos that present Craig in the best possible light. Once you have them all at hand, you can attach them to a separate collection and in 2 clicks create a website that you can share with anyone you need!
NB! Attaching assets to collections is different from moving. You can attach one asset to multiple collections and they will exist in multiple places at once. It does not mean that the asset will be duplicated (so no storage costs) but that it can be looked at from multiple places. This is useful for Websites in particular where you want to showcase some assets but don't just want to ruin carefully constructed folder and asset structure in the process.
Keeping Track of Licensing and Model Releases
Besides photos, you can also upload documents - licenses and model releases - to your Pics.io. You can tag them with same keywords that you would photo assets but you can also use the linked assets feature to connect photos processed by facial recognition solution and these documents like so:
Returning to our imaginary friend Craig here. Once you've recognized all photos with Craig, you can then bulk link those photos with whatever model releases they might have, and quickly hop over to their contract to see what's going on there.
Face recognition is a great addition to the digital asset management toolkit, as the software's main purpose is to help you keep track of all assets that you have and use them in a meaningful way. With face recognition capabilities, teams that work with photos of people on a daily basis get a powerful feature that lets them use face recognition on bulk imagery and then search and manipulate those assets as they fit.