The neural network will find similarities between the user’s photographs and world masterpieces.

Twitter

Perhaps you have long noticed that you look like someone from a movie, but you still haven’t had the opportunity to find out who exactly it is. Or maybe you've always wanted a doppelganger. With the development of the Internet, many free services that bring people together based on their interests. Today there are several different sites that will help you find your photo match online.

Pictriev.com – online service for searching for doubles by photo without registration

People have always wanted to be like celebrities. Girls especially strive for this, who often find something in common with Angelina Jolie, Pamela Anderson, Marilyn Monroe and always emphasize these similarities. Technologies of the pictriev.com service will help you quickly find the same facial features not only among celebrities, but also among others ordinary users. To do this you must have high quality photo in a full-face view, in which it is desirable that there be a gap between the eyes of 80 pixels. The minimum approximate photo width is 200 pixels. The image weight should not exceed 200 KB. Using the service is quite simple.

For this:

How to find doppelgangers in social networking groups

To date software has reached its peak of development. Specially trained robots can play football and even stand on the goal and parry the ball! But before human brain they still have a long way to go. One of the main and incomprehensible human feelings for a computer remains intuition.

In the most extensive online databases photographs - on social networks there are special groups that search for doubles. Here you are most likely to find a person very similar to you, because the search is carried out not only software, but also by other people. You can find such a group in almost every Russian-language social network. For a successful search, it is advisable to have a genuine and high-quality photo as your avatar, where your face will be photographed as closely as possible.

How to find such groups and start searching:

Open a search on VKontakte or Odnoklassniki and enter “Double”.
A window with similar groups will open, select the most popular one and join it.
In the group you can find information about what needs to be done to find a double.

One more effective way is the search for a double in search engines using photographs of other search engines. After all, unlike social networks, search engines search through all resources at once, including all social networks on the Internet. This means that finding a person similar to yourself is very high. two will help you popular search engines– Yandex and Google. For a successful search, you will need a high-quality photo of you. You can make it right now and download it to your computer, putting it in a folder where you can easily find it.

The search itself will not take much time, for this:

Similar actions must be carried out in the Yandex system to find your double:

FindFace - quickly search for people using photos

You've probably had such cases when you had to search for a person on social networks based on some data that you had. Usually there is very little of this data, and finding a specific user was exactly what was needed to get more information on it. And to find someone, I had to use various filters - country, city, school, age. With the advent of the new FindFace service, you can easily find your doubles by photo. To do this, you just need to upload a file with your image and the service will automatically search by facial features.

The technology is built on the use of neural networks, which are independently improved and each time the results become more accurate. The service searches its database among 300 million photos in less than one second. This becomes possible thanks to special technology in which the photo file is converted into a special internal representation of characters. This is what gives quick and accurate results. FindFace works with all platforms - Windows, iOS, Android.

You can find a double of any person; to do this, you just need to take a photo of him and upload the photo to the service. FindFace searches all popular social networks. You can use the service for free, but there are some restrictions. For free subscription 30 uses per month available. If you have exhausted all attempts, you must purchase paid subscription. The price for premium is 149, the number of uses per month is 75 times, and there are no ads. VIP price – 375 rubles, 300 times of use, no advertising and other bonuses.

Using the service is very simple:

Assessing their appearance in the mirror, many girls sometimes try to understand which world cinema star they might resemble. Catherine Zeta-Jones or Jennifer Lopez, Julia Roberts or Angelina Jolie, Kristen Stewart or Nicole Kidman - which celebrity is exactly like you? Suggesting that you do not guess on the tea leaves, but turn to the help of network services that, based on neural networks and various digital algorithms, will give a suitable answer. In this article, I will list services that can determine your resemblance to a celebrity from a photo online. I also suggest using the service to search among all people.

There are a number of resources on the Internet (especially in the English-speaking segment) that allow you to determine your similarity with a celebrity online. Working with them is quite simple - you launch such a resource, upload your photo to it (or provide a link to it on the Internet), and click on the verification activation button. A verification procedure is performed, and then you receive the result in the form of a photo of a celebrity similar to you and the percentage of similarities identified.

At the same time, some of these resources may provide other information on the photo being processed (the gender of the person in the photo, his age, etc.), which, of course, should not be taken seriously.

Let's move on to the list of services that allow us to figure out what kind of celebrity I look like online.

Celebslike.me – will select the celebrity most similar to you

The celebslike.me service is the most popular network resource for determining similarity with a star online. It was created by Microsoft for those users who would like to understand which Oscar nominee they might be like.

To start the search procedure, you need to go to celebslike.me.
Click on the “USE YOUR OWN PHOTO” button and upload your full-face photo to the resource.
Literally immediately you will receive a photo of an Oscar-nominated celebrity similar to you (with a percentage of similarity), as well as three more celebrities below who are not nominees.

Another Russian-language service that allows you to identify which celebrity you resemble is pictriev.com. Feature online resource is to find not only a movie star who looks like you, but also to determine the gender and age of the person in the photo. Of course, you shouldn’t expect of this algorithm accuracy of definition, take it as entertainment, in a humorous way.

To work with the service, log in to pictriev.com.
Click on “Upload image” and upload your photo to the resource (frontal photos up to 200 kilobytes are accepted).
In a few seconds you will receive information about the found similar celebrities(with the percentage of similarities), the estimated gender and age of the person in the photo.

Celebritymatch.io – perfectly identifies photos

The English-language resource celebritymatch.io was written by the author in the Wolfram programming language and is designed to find a celebrity similar to you. Unlike other resources, it gives only one star similar to you with a percentage of similarity (there are more than 2 thousand stars in the resource base).

To use the resource, go to celebritymatch.io.
And click on the “drop a photo here to find your celebrity match” button.
Upload your photo to the resource, and in a couple of seconds you will receive the name and photo of the star most similar to you.

En.vonvon.me – determine similarity using a Facebook profile

There are several entertainment services that work based on analysis of a user’s profile on Facebook (for example, ru.namestests.com, popular on the Runet). I suggest using another popular service vonvon.me, which will analyze your Facebook photo and tell you which star you resemble online.

Go to en.vonvon.me and select “Continue with Facebook” (you will need to give the service permission to analyze your public profile).
The service will analyze your photos and, if necessary, ask which photo from those available in your profile to choose, as well as your gender (male - male, female - female).
Then you will get your result.

Search by photo from Facebook profiles"Vonvon"

Playbuzz.com - entertaining search in the form of answers to questions

The Kate Walker test posted on playbuzz.com will ask you to answer 6 basic questions about your facial features. Based on your answers, the system will select the celebrity who is most visually close to you.

Go to playbuzz.com and click on the “Let’s play” button.
In the first question, select the color of your eyes and hair color.
Features of hairstyle (hair style), skin color (skin tone), shape of nose (nose), lips (lips), and face (face).
At the end you will get your result.

Complementary mobile applications

In addition to the network services listed above, it is worth using various mobile applications to determine similarity to a star. Among such applications, I would highlight the following:

Conclusion

Listed above network resources and mobile apps will let you find out which celebrities your photo resembles. The results obtained with their help should not be taken too seriously - this is just fun entertainment, and nothing more. Happy coincidences!

In contact with

How does the procedure for searching for a double among stars work?

Let's move on to the list of services that allow us to figure out what kind of celebrity I look like online.

Celebslike.me – will select the celebrity most similar to you

To start the search procedure, you need to go to celebslike.me.
Click on the “USE YOUR OWN PHOTO” button and upload your full-face photo to the resource.
Literally immediately you will receive a photo of an Oscar-nominated celebrity similar to you (with a percentage of similarity), as well as three more celebrities below who are not nominees.

Another Russian-language service that allows you to identify which celebrity you resemble is pictriev.com. A special feature of the online resource is that you can find not only a movie star who looks like you, but also determine the gender and age of the person in the photo. Of course, you should not expect accurate determination from this algorithm; take it as entertainment, in a humorous way.

To work with the service, log in to pictriev.com.
Click on “Upload image” and upload your photo to the resource (frontal photos up to 200 kilobytes are accepted).
In a few seconds you will receive information about the similar celebrities found (with the percentage of similarities), the estimated gender and age of the person in the photo.

Celebritymatch.io – perfectly identifies photos

To use the resource, go to celebritymatch.io.
And click on the “drop a photo here to find your celebrity match” button.
Upload your photo to the resource, and in a couple of seconds you will receive the name and photo of the star most similar to you.

En.vonvon.me – determine similarity using a Facebook profile

There are several entertainment services that work based on analysis of a user’s profile on Facebook (for example, ru.namestests.com, popular on the Runet). I suggest using another popular service, vonvon.me, which will analyze your photo on Facebook and tell you which star you resemble online.

Go to en.vonvon.me and select “Continue with Facebook” (you will need to give the service permission to analyze your public profile).
The service will analyze your photos and, if necessary, ask which photo from those available in your profile to choose, as well as your gender (male - male, female - female).
Then you will get your result.

Search by photo from Facebook profiles “Vonvon”

Playbuzz.com - entertaining search in the form of answers to questions

Go to playbuzz.com and click on the “Let’s play” button.
In the first question, select the color of your eyes and hair color.
Features of hairstyle (hair style), skin color (skin tone), shape of nose (nose), lips (lips), and face (face).
At the end you will get your result.

Complementary mobile applications

In addition to the network services listed above, it is worth using various mobile applications to determine your similarity to a star. Among such applications, I would highlight the following:

Conclusion

The online resources and mobile applications listed above will allow you to find out which celebrity your photo resembles. The results obtained with their help should not be taken too seriously - this is just fun entertainment, and nothing more. Happy coincidences!

Good afternoon In today's article we will talk about finding a double using a photo for free online. Finding your double using a photograph is very interesting task. Unfortunately, the probability of meeting a “twin brother” on the street tends to zero.

Modern information Technology they can help us find our double using facial recognition programs right while sitting on the couch. Some people don’t even realize that they look like a celebrity or just another person. Most of the proposed methods are ineffective and do not lead to positive result. In this case, using sites and online service with an already proven reputation.

Necessary conditions to start searching:

laptop or computer with Internet connection;
photograph with which comparison will take place.

In addition, the direction of a person’s gaze in photographs is the key to a successful search. If a person's gaze is directed to the side, the result may not be accurate. Looking forward will increase the chance of a successful search. A plain background will also contribute to obtaining a high-quality result. It is enough that the photo only contains an image of the face; there is no need to upload a full-length photo.

How to find a double among famous people?

It's interesting to find a person among celebrities who looks like you. This question haunts many people. To accomplish this task, there are many free, proven sites with online search capabilities.

Play-analogia.com is a site that allows you to find 3 stellar doubles. If I want to use the search on this site, I must register. After that, upload the photo (don’t forget about the important conditions when choosing it), indicate female or male gender. To start the search, you need to mark the points in the center of the pupils in the photo. Click the “Send” button and get the result. If the search is unsuccessful, the procedure can be repeated using the “Again” button.
Myheritage.com is a resource containing about 4,000 photographs of celebrities. The procedure is similar: download the application, register and start searching.

Online services for lookalikes

The most famous online services are:

Twins (http://ts.family-site.com.ua) – popular network program to search for doubles. Registration is not required to access the resource. The search process itself takes about 6 minutes. On the loaded image, the program independently detects about 50 significant points and conducts a similarity analysis based on them.
Stranger Twins (http://twinstrangers.net). The search takes 5 minutes from the moment the photo is uploaded. You are asked to choose the facial features that will be used for comparison yourself, unlike Twins. The website also indicates the criteria that the photo must meet.

Search order on the PicTriev.com website

This resource is quite simple and effective. In order to use the search, you need to perform the following steps:

Open the PicTriev.com website page. Select “Upload image”;
Select a photo to search. Don’t forget about the conditions for choosing a photo to receive best result search.
The result will appear in literally 10 seconds. The program will determine the age, gender and names of stars with whom there will be similarities. You can immediately determine how correctly a given resource works.

The service is popular among people who have recently changed their appearance and want to find out the percentage of similarity with a celebrity.

This resource works great with photographs of women. Screenshot good example You can see such a search below.

Search for a double by photo for free online in social networking communities

The VKontakte and Odnoklassniki websites will help you find twins.

Group "Doubles" ( http://vk.com/club15522443). Users share their photos, opinions, and look for similar people.
Find a double ( http://ok.ru/group/51792531357844) – search for doubles in the database, posting your photo.

Searching for a double through systems on the Internet

The most accurate result can be obtained using various search engines. The complication of algorithms and the increase in the amount of loaded information day by day makes this search option better and more effective. As an example, consider the Google search order.

Open home page http://www.google.ru;
On the right top corner click on “Pictures”;

In the search bar you will be able to search by images. Select the "Camera" icon;

Neural networks have reached art. Thus, thanks to them, the project “Find yourself in art” was recently created, implemented by the R.Point agency and the NTechLab company for the 175th anniversary of Sberbank. On the project website, anyone can upload their photo and find their “double” in world portraiture. the site figured out how this technology works and how a neural network can “see through” even a masked criminal.

The “Find Yourself in Art” project uses a neural network created by programmers from the NTechLab company. The company's development was recognized as the best facial recognition algorithm according to the world championship The MegaFace Benchmark, held at the University of Washington. the site spoke with NTechLab co-founder Artem Kukharenko and found out exactly how a neural network works, what else it can do, and whether it can be deceived artificial intelligence.

How does the facial recognition system used in the “Find Yourself in Art” project work? Is this artificial intelligence based on neural networks? If so, what are the definitions of the terms “artificial intelligence” and “neural network”?

The “Find yourself in art” project was implemented using NТеchLab photo facial recognition technology. Our company creates solutions using advanced methods in the field of artificial neural networks and machine learning.

An artificial neural network or "artificial intelligence" is a mathematical software package, built on the principle of organization and functioning of biological neural networks - networks of nerve cells of a living organism.

Example of how the algorithm works

Yulia Alekseeva/website

Neural network provides a set of characteristics by which one can be distinguished from another (eye color and shape, facial expressions, etc.). But most of the signs that a neural network produces are invisible to the human eye. The accuracy of image determination by a neural network is about 90%, and by a person - 25% (with a database volume of, for example, ten thousand photographs).

- What exactly is the NTechLab algorithm capable of?

Our algorithm makes it possible to compare pairs of faces with a 99% degree of accuracy and search through a fairly large database of photos in less than 0.3 seconds with an accuracy of more than 70%. This technology was recognized as the best at the world championship The MegaFace Challenge, organized by the University of Washington in 2015. More than a hundred teams from all over the world took part in this championship, including the Google team.

To search for a person in a database of one billion photos, such an algorithm will take less than a second. Such a search speed can solve many problems not only on a city scale, but on a national and even global scale. For example, searching for a criminal in real time.

Among the advantages, in addition to the speed of searching through global photo databases, the algorithm has very high recognition accuracy. This became possible thanks to deep learning and properly selected neural network architecture.

- How exactly does image recognition occur?

The recognition process is the construction of a feature vector using a trained neural network. The feature vector consists of 80 numbers that contain all the information about the face. For one person the numbers are similar, for two different people they are different. The search system is built on this difference. It is important that facial information does not change if a person puts on glasses, grows a beard and mustache, or if there is a difference of several years between photographs.

It is worth noting that information about one person takes up less than a kilobyte for each image, which allows the algorithm to operate without using large computing power.

- Is it possible to fool the recognition system by wearing a mask?

Our algorithm is quite resistant to changes in appearance and is able to identify people if they wear glasses, grow a beard, or change their hairstyle. You can recognize a person even in a medical mask, but the percentage of accuracy will be lower. We are working to ensure that the drop in recognition accuracy in such situations is minimal.

In that case, could you give examples of the features that the neural network pays attention to? After all, if a person is wearing a medical mask, then a significant part of the face will be hidden.

During the learning process, the neural network itself generates features and selects them depending on statistical patterns in the data. They are usually not interpretable by humans.

- What size is the database that NTechLab has?

Is it possible to use this technology in the field of law enforcement? For example, to recognize the faces of criminals in videos from security cameras or in photographs on social networks?

Yes, sure. Face recognition algorithm can be widely applied in various fields such as retail, banking services, security, entertainment industry, sporting events, dating services and many others.

- Are there any plans to create mobile applications based on this technology? If yes, which ones exactly?

Based on the NTechLab algorithm, FindFace, a service for searching people by photo on the VKontakte social network, has already been launched, which gained more than a million users in the first three months. The service was launched on February 18, 2016 as a demonstrator of our technology. To date, we have launched two products for business - cloud FindFace service Cloud API and the FindFace Enterprise Server SDK solution. Now any company in the world can integrate NTechLab facial recognition technology into its operations. The service offers two main work scenarios: verification (comparing pairs of faces) and identification (searching for faces) using its own photo databases of any scale.

- Google recently announced a game based on a neural network that recognizes the user’s drawings. Is this algorithm similar to the development used in the “Find Yourself in Art” project? Are these competing developments? Or, by combining them, you can get artificial intelligence with superpowers?

Neural networks can be trained under absolutely different tasks: recognition of images, speech, writing, objects, emotions and much more. Many global corporations, such as Google, Facebook, etc., are working on these issues. Face recognition is one of the most difficult tasks. To date, no algorithm has been able to demonstrate such speed and accuracy of searching through global databases.

About NTechLab

Founded in 2015 by Artem Kukharenko and Alexander Kabakov, the startup NTechLab specializes in advanced methods in the field of self-learning neural networks for creating facial recognition algorithms.

In February 2016, based on the company’s algorithm, the FindFace service was launched to search for people in social network“VKontakte” based on photography, whose audience in less than three months exceeded a million people. At the beginning of August, the NTechLab solution was used at the largest electronic music festival in Russia, Alfa Future People, which was held with the support of Alfa Bank and VimpelCom. More than 15 thousand people used the service. In September of the same year, NTechLab was selected as a supplier of facial recognition solutions by Diamond Fortune Holdings Prim, which will build hotel and entertainment complexes on the territory of the Primorye integrated resort. In October NTechLab launched cloud service for business FindFace.PRO, with the help of which any company around the world can search its own database of any scale. And already in November 2016, Bitrix24 integrated it into its CRM system, becoming the first to introduce facial recognition into retail. In the same month, NTechLab launched FindFace Enterprise Server SDK.

According to an influential American portal in the field high technology Techrepublic.com, NTechLab is one of the five most attractive Russian hi-tech startups for potential investors.

The neural network will find similarities between the user’s photographs and world masterpieces.

Pictriev.com – online service for searching for doubles by photo without registration

How to find doppelgangers in social networking groups

FindFace - quickly search for people using photos

Celebslike.me – will select the celebrity most similar to you

Celebritymatch.io – perfectly identifies photos

En.vonvon.me – determine similarity using a Facebook profile

Playbuzz.com - entertaining search in the form of answers to questions

Complementary mobile applications

Conclusion

How does the procedure for searching for a double among stars work?

Celebslike.me – will select the celebrity most similar to you

Celebritymatch.io – perfectly identifies photos

En.vonvon.me – determine similarity using a Facebook profile

Playbuzz.com - entertaining search in the form of answers to questions

Complementary mobile applications

Conclusion

How to find a double among famous people?

Online services for lookalikes

Search order on the PicTriev.com website

Search for a double by photo for free online in social networking communities

Searching for a double through systems on the Internet

Popular articles

Latest articles

Sections

Pages

Special projects

Contacts