How to Use Facial Recognition Search
Facial recognition search technology has come a long way in recent years. Combined with facial identification technology, face recognition technology not only allows software to recognize when a face occurs in an image, but to also identify other images with the same face or even confirm an identity in real-time.
The applications for this kind of tech are vast – from checking to see if anyone is using an image illegally to unlocking your phone.
For photographers, it does everything from in-camera face detection to instantly pulling up every image in your library pertaining to a particular person.
If you’re an event or portrait photographer, automatic face tagging can save you incredible amounts of time. For hobbyists who love family photos, facial recognition technology can make finding photos of loved ones infinitely easier.
When used in a reverse image search, face recognition can also be used to see if anyone has taken your images without permission, violating copyright.
Let’s take a look at how it’s done and how best to use it.
How is Face Recognition Done?
There are two different types of facial recognition – fiducial recognition and pixel detection.
In fiducial recognition, the software measures the dimensions of different facial features. Examples include the distance between the eyes, the width of the nose, the length of the jawline, or even the shape of the cheekbones.
Pixel detection analyzes the pixels in an image and compares them to a database of appearance generalizations (i.e. age, sex, etc.).
Both methods collect biometric data from a person’s face and facial expression and use this to identify and/or verify the identity of the person in the photo.
Paired with artificial intelligence and machine learning, face recognition technology has become more and more accurate, and now can even work with incredible accuracy in real-time – at least in certain demographics.
What is a Reverse Image Search?
A reverse image search is the process of uploading a photo into a search engine and directing it to find similar images. This can be anything from people to car parts to plant IDs.
The search results will not only show similar images but also websites that have that image or ones similar to it.
At its most creative, reverse image search turns into a sort of Shazam for images, except that instead of songs you’ll get everything from Wikipedia articles to links to relevant websites to prices and places to buy an item.
In terms of facial recognition, reverse image search was originally used for entertainment purposes, like finding the name of a celebrity you may have forgotten.
These days, however, there are all sorts of applications – everything from discovering fake accounts on social media sites to discovering copyright violations.
You can even use it to find out how far into the wide reaches of the internet images of your own face have traveled!
How Can you do a Facial Recognition Search?
Interested in doing a face search online? There are a number of services that offer a basic face search for free. For a more refined result, though, you’ll probably have to invest a little, depending on how deep into the Internet you want to go.
Here are a few of the more popular face finder options.
One of the newer face recognition search engines out there, PimEyes searches for similar faces on over 10 million websites. They offer both free and paid options.
The free option only allows one to see if a certain face is on the Internet. The subscription service allows you to access all of its other services, including Deep Search, generating PDFs, and sending you notifications if your photos come up later.
PimEyes can even help you get your image removed from a site.
One cool thing about the results is that they’re not limited to the photo you uploaded. The service also finds pictures where the background’s been changed, other people have been added, or modifications have been made like a change in hair color.
Their search results don’t find everything, though. I tried it with a couple of stock photos. On one search they didn’t find anything. On the other, they found several.
- Is PimEyes legal?
Yes, PimEyes is legal.
- Does PimEyes keep your photo?
PimEyes saves your photo for 48 hours. After that, it is expunged. It does not store photos or any private data from our users.
- Does PimEyes save your face?
PimEyes does not store any images. It only displays images from source websites.
Perhaps the most well-known reverse image search engine, Google Images allows you to reverse search just about anything, including faces.
To get started, click the camera icon in the search bar. You can either upload a photo or paste the image URL. Google will then find similar images.
Once your results load, click on the image and then go to Tools to pull up more search options. Under the Type pulldown menu you should see an option for Face. Select that and you’re good to go.
In addition to going to the Google Images search page through a browser, you can also access the service through a number of different mobile apps.
For example, the Image Recognition and Searcher app will reverse image search Google, Yandex, and Bing, among others. (Available for iOS 0nly.)
Perhaps my favorite mobile reverse search engine app, Google Goggles not only does faces, but it also accesses information about historical landmarks, let you scan barcodes, scan business cards, translate text, and even solve Sodoku puzzles.
If you use an Android phone, you can also use Google’s face search abilities in Google Photos.
If you’re a fan of looking up celebrities or would like to figure out which celebrity you look most similar to, check out PicTriev.
Just make sure whatever photo upload you do is no larger that 200 KB.
One bonus feature of PicTriev is that it will also let you estimate whether two different photos are the same person.
Betaface is similar to PicTriev but goes beyond looking for celebrity faces. Like the search engines above, you either do an image upload or use an image’s URL.
Once it has a photo (or two) to work with, you can then compare faces, search Wikipedia, or search an individual website. You can even create your own user database for Betaface to search.
Beyond these, there are a few other services out there, but the results definitely get a bit murkier. It’s definitely best to try these four first!
Pros and Cons of Facial Recognition Technology
As mentioned before, face recognition has a number of amazing applications all across the board. It also, however, has a number of downsides.
Many are already using it to unlock their phones or to tag their friends in Facebook posts. It also has a number of benefits for those in the photography business.
For photographers using Adobe Lightroom, you can use Adobe’s excellent face recognition abilities to organize your photos of people in its Library – this function is currently only available in the Cloud-based Lightroom as opposed to the desktop Lightroom Classic app.
Face detection also enables our autofocus to quickly focus on the eyes in a portrait or an event photo, ensuring that they’re in focus.
Online, you can check to see if anyone is illegally using either your personal photo or those you’ve copyrighted.
Street photographers can reverse image search for people in their shots if they’d like to transition a shot to a stock photo and get a model consent form.
Unfortunately, the use of face recognition technology comes with a number of alarming downsides.
Most of us don’t really want our personal images recorded and then stored in a government or corporate database for unknown future use.
Law enforcement, for example, can use facial recognition to run anyone in their database through a virtual criminal lineup, which is pretty much the same as treating you as a criminal suspect, but without probable cause.
This issue is so large that a number of US cities have banned the use of real-time (live) facial recognition surveillance by law enforcement – Cambridge, MA and San Francisco, CA among them.
On a more personal level, identity thieves can use a facial recognition search to collect not only personal information but also find images and videos of you to aid them in creating everything from false social media accounts to fake IDs. It can also be used to spread revenge porn.
There have also been reports of some companies, in an effort to design the next, best face finder, snagging thousands of face images off of sites like Flickr, and feeding them into their databases without the consent of the photographers or models.
Another issue with face detection technology is that it’s still far from perfect when a face search includes anyone other than a white male.
Women and people of color are especially problematic, as the algorithms being used have far more data on white men than anyone else, creating a definitive bias.
This can be particularly problematic in terms of law enforcement, as a false positive can often lead to an arrest.
While there’s definite room for improvement, facial recognition search and reverse image search technology has already come a very long way.
Combined with other technologies – for example, the AI of On1 and Luminar or the fantastic Library module in Lightroom, it’s a fantastic tool for photographers who focus mostly on shooting people.
The reverse image search feature of some search engines is quite likely the way of the future, both for people and for objects and places. (Check out Google Lens for an excellent reverse image search engine for objects and landmarks!)
And privacy? Well, for good or for ill, facial recognition technology is only going to get better. If you’re really worried about how your personal image will be used, it’s best to keep your privacy settings as high as possible when posting or sharing online.
What do you think? Is privacy an issue? Do you use facial recognition technology, either in your photography or elsewhere?
Usnea Lebendig is a travel and landscape photographer who loves trekking in the wilderness, exploring other cultures, and using photography for social activism.