TikTok is a popular short-form video app known for its fun effects and filters that allow users to alter their appearance. In 2021, a filter called the “Who Do You Look Like” filter went viral on TikTok. This filter uses facial recognition technology to analyze a person’s face and tell them which celebrity they most closely resemble.
The “Who Do You Look Like” filter sparked a huge trend on TikTok, with countless users trying it out and sharing their results. Videos tagged with #CelebrityLookAlike have over 3 billion views on TikTok. The filter’s popularity stems from people’s curiosity about which stars they look similar to. It’s an entertaining way for users to see themselves in a new light.
What is the TikTok ‘Who Do You Look Like’ Filter?
The ‘Who Do You Look Like’ filter is a popular TikTok filter that shows users which celebrity they most closely resemble. The filter uses facial recognition technology to analyze a user’s facial features and match them against a database of celebrity photos.
According to an article by The Tab, the filter first launched on TikTok in early 2023. When a user activates the filter, it scans their face and generates four celebrity lookalikes based on their bone structure, features, and facial contours. The AI compares the user’s selfie to thousands of celebrity images to find the closest matches.
The filter went viral in early 2023, with many TikTokers posting reaction videos of their celebrity doppelgangers. Some were delighted at who they were matched with, while others poked fun at more unflattering comparisons.
How the Filter Became Popular
The ‘Who Do You Look Like’ filter on TikTok went viral in 2022 after numerous videos using the filter started trending on the platform. Some of the most popular videos featured people shocked at seeing which celebrity they were matched with by the filter, like TikTok user @gabbixsmith who exclaimed “Who is this person??? It’s crazy that we will never know” when shown her celebrity lookalike (1).
Seeing funny and surprising reactions to the filter in these viral videos inspired many others to try it out. Celebrities also started jumping in on the trend, posting videos of themselves trying out the filter. Actor Will Smith used the filter and joked “Who do you think I will look like?” before being matched with a photo of himself (2). The excitement and buzz surrounding the viral videos and celebrity attempts with the filter helped it gain immense popularity on TikTok.
(1) https://www.tiktok.com/@gabbixsmith/video/7145999039018913070
(2) https://www.tiktok.com/discover/who-do-i-look-like-filter-on-tiktok?lang=en
Trying the Filter as an Average User
Downloading the “Who Do You Look Like” filter on TikTok is easy. Simply open the TikTok app, go to the Discover page, and search for “Who Do You Look Like.” Once you find the effect, tap on it to download it to your own account.
To try out the filter, go to the camera in the TikTok app and select the “Who Do You Look Like” effect. Position your face in the viewfinder and snap a selfie. In a few seconds, the filter will shift and show a celebrity’s face over yours that you apparently resemble. The results can be hit or miss, but also hilarious and entertaining.
According to a test by In The Know, the filter matched an average male user with comedic actor Seth Rogen. For an average female user, the filter thought she looked like Marilyn Monroe. But again, your mileage may vary depending on lighting, angles, and other factors. Celebrities like Emma Chamberlain, Charli D’Amelio, and Addison Rae have all tried the filter too out of curiosity.
While the filter results aren’t always accurate, the “Who Do You Look Like” effect has become hugely popular for entertainment. It’s fun to try with friends and see which celebrities you each get matched with. The TikTok effect has clearly resonated by letting users see their celebrity doppelgängers.
Behind the Scenes
TikTok develops filters like the ‘Who Do You Look Like’ filter using advanced computer vision and machine learning technologies. According to Effect House, TikTok’s filter creation platform, filters are powered by OpenGL, OpenGL ES, and Metal, which are graphics libraries that render the visual effects in real-time. The computer vision algorithms analyze the user’s facial features like eyes, nose, mouth, etc. and compare them against a database of celebrity faces to find the closest match. The matching process likely uses deep learning and neural networks trained on massive datasets of celebrity photos.
To develop a filter from scratch, TikTok engineers use Effect House to combine different components like face tracking, segmentation, eyelash effects, lenses etc. The face tracking identifies facial landmarks, the segmentation separates the foreground subject from the background, and the AR effects overlay virtual elements like eyelashes on the person’s face in real-time. TikTok engineers can tweak the filter’s code to improve the accuracy of the computer vision and fine-tune the lookalike matches. Overall, filters like “Who Do You Look Like” demonstrate TikTok’s advanced computer vision capabilities and blending of entertainment with AI technology.
Impact and Criticism
The “Who Do You Look Like” filter has sparked debate around its potential negative impacts, especially for young people. One major concern is the promotion of unrealistic beauty standards. By showing idealized versions of users, filters like this imply there is one narrow definition of attractiveness we should all aspire to. This can potentially harm self-esteem and body image, particularly among adolescents. According to one source, “Such filters set unrealistic expectations that no one can healthily live up to in reality.”
There are also fears about privacy violations. To create detailed facial recognition capabilities, these filters gather enormous amounts of personal data. However, consent around data collection is often unclear. As reported by one analysis, “What information gets stored, and for how long, remains very opaque.” This has led to calls for greater transparency and regulations around data usage. Overall, lookalike filters demonstrate some of the emerging ethical challenges as facial analysis technology becomes more advanced and integrated into daily apps.
Celebrity Lookalikes
One of the most popular uses of the TikTok “Who Do You Look Like” filter is seeing which celebrity you resemble. Many fans have been delighted to find that they look like their favorite actor, musician or influencer. According to TikTok, some of the most common celebrity matches from the filter include:
Zendaya – With her striking features and fashionable style, many users are thrilled when the filter matches them with this Emmy-winning actress.
Timothée Chalamet – This popular young actor often comes up as a top match, likely due to his curly hair and chiseled jawline being common traits.
Billie Eilish – From her green hair to her unique fashion sense, lots of fans find they resemble this Grammy-winning singer.
Fans often post reaction videos of themselves delightedly gaping at the screen when they’re matched with their celebrity twin. Part of the fun for users is seeing if the algorithm agrees with who they think they look like. It’s become a trend to duet with the celebrity you’re matched with, showing side-by-side how similar you look.
Tips for the Best Results
Here are some tips to get the most accurate and best results from the TikTok “Who Do You Look Like” filter:
Lighting – Make sure you have soft, even lighting on your face. Side lighting or harsh shadows can distort your facial features. Natural daylight often works best.
Angles – Hold your camera at eye level or slightly above. Pointing the camera up at your face can make features look wider, while pointing down can make them appear narrower.
Expressions – Keep a neutral facial expression, relaxing your face. Smiling or making exaggerated faces can change the way your facial structure looks.
Distance – Frame your face so it takes up most of the screen without being cut off. Being too close or far away can decrease accuracy.
Troubleshooting – If you aren’t getting good results, try adjusting the camera angle, zooming in or out, or using a different expression. Good lighting also helps.
Take multiple tries – Slight changes in angles, lighting, and expressions can produce different results. Take a few snaps to see the variety of celebrities you can get.
Reference: https://www.tiktok.com/discover/what-you-really-look-like-filter?lang=en
Other ‘Who Do You Look Like’ Apps
In addition to TikTok’s built-in filter, there are several apps dedicated to finding celebrity lookalikes. Some popular options include:
Celebs – Celebrity Look Alike – This iOS app has a database of over 100,000 celebrity photos and uses facial recognition to find the closest matches. Users can upload a selfie or use a live camera to find their lookalikes. The app is free but limits the number of searches per day without an in-app purchase.
Gradient – This app went viral in 2019 for its lookalike filters and comparisons to other celebrities. It uses AI for facial analysis instead of a direct database. The app is free on iOS and Android but also has in-app purchases. While popular, some reviews suggest the results are not as accurate as dedicated celebrity lookalike apps.
Compared to Gradient, apps like Celebs likely have more accurate results since they directly compare user photos against a large database of celebrity images. However, TikTok’s filter has the advantage of being built into the platform so it can spread virally. Dedicated lookalike apps can provide a more robust experience for users specifically interested in finding their celebrity doppelgängers.
The Future of Lookalike Filters
Lookalike filters rely on facial recognition technology, which is rapidly improving. In the future, these filters may become even more accurate at identifying celebrity doppelgangers. The algorithms powering the filters will likely continue to advance, analyzing facial features and proportions in greater detail.
As the technology progresses, lookalike filters could gain abilities like morphing users’ faces in real-time to look more like a chosen celebrity, or generating hybrid faces blending the user’s appearance with a star’s. Filters might also expand beyond celebrities, allowing users to see their similarity to historical figures, animals, cartoon characters and more.
New viral filters may emerge that judge resemblance based on users’ hairstyles, dressing style, makeup or other characteristics beyond facial features. The virality of the ‘Who Do You Look Like’ filter demonstrates people’s fascination with seeing altered versions of themselves. As computer vision and generative AI advance, we can expect more sophisticated filters that transform users’ appearances in entertaining ways while generating viral buzz.
