Who Do You Look Like

"Find Your Doppelganger in the Stars!!"

Comment

Inspiration

We were inspired to create this project because we believe that everyone has a unique and special quality that makes them stand out. And sometimes, people are compared to a celebrity because of their facial features, smile, or even their personality traits. With this project, we aim to bring this comparison to life by using technology to match individuals with their celebrity lookalikes. Our goal is to provide a fun and engaging experience for users to discover the celebrity they resemble, and to celebrate the individuality and uniqueness of each person. We hope that this project will bring joy and be a fun thing to play with.

What it does

Discover the celebrity you have always been compared to and uncover the famous face behind your own. With cutting-edge technology and advanced facial recognition algorithms, our project, "Who You You Look Like", brings you closer to finding your true celebrity twin. Embrace your inner star and let the world see the famous reflection in the mirror. Let's start the journey of self-discovery and see which celebrity you resemble the most.

How we built it

Firstly, we collect images of many popular actors and collate them in separate folders. We then use OpenCv's face detector to detect the faces in each of these images. Next we use Support Vector Machines to determine embeddings for these actors. Once that's done, we take the user input image and detect the face in the image and calculate the embeddings. Finally we compare the embeddings and give out a confidence score or in short match score of your image with the images of these actors.

Challenges we ran into

Due to less time and non availability of resources, we could only collect images for around 18 actors. This project can be extended to a huge database as well in the future.

Accomplishments that we're proud of

Originally, the run time for detecting was around 3 minutes. We optimised our code along with the dataset to reduce this runtime to around 10 seconds. This is something that we are extremely proud of!

What we learned

While building this project, we learned a lot about data collection. Along with that, trying to fix small hiccups within the given time, made us aware of all minute things that we should take care of while builiding such projects.

What's next for Who Do You Look Like

The next thing for this project would be to increase the database of actors. We could also put in a feedback system so that users can agree or disagree with the celebrity matches we produce, hence improving our system!

Built With

Share this project:

Updates