HOW HOT AM I?
Test it now by uploading a picture of you. The beauty test rates your face on a scale from 1 to 10.
Recommendations for accurate beauty-predictions
How should the picture look like? Here are Examples that work for calculation the attractiveness score.
Besides these tips, we also recommend avoiding accessories like sunglasses and nothing should cover parts of your face. Furthermore, we suggest not using black and white pictures and no Instagram filters that drastically change the colors. You should also have a neutral or smiling face expression. Note that hair doesn't have a significant impact on the beauty test, much more important is your face. We recommend you stick to these suggestions so that our AI-Model can return an accurate score.
Note that how attractive you appear to others can be influenced and does not depend only on your look. Read here how you can become more attractive for other people.
Articles in our Blog
We try to regularly add new articles in our blog about attractiveness and how you can look and feel better. Some articles redirect to our partner website that focuses on physical health.
Example Results of the Beauty Rater
The following photos were all rated by our beauty calculator. The people in the photos are not real and are all created by Generative Adversarial Neural Networks. You can find more information here. a scale from 1 to 10. The pictures are sorted from less attractive to very attractive.
Female faces ordered from least hot to hottest
Male faces ordered from least hot to hottest
Technical details about the AI-model
We use a customized deep learning model, based on a pre-trained ResNet-50 architecture which is finetuned with a dataset of hand-labeled photos.
A second Neural Network locates the faces, crops it, and transforms a bit, to make the input images consistent with our training dataset. We also use Open-CV for the initial preprocessing. We plan to publish a more detailed technical description at a later time as a blog post. This application is the outcome of a student project about deep learning and computer vision.
Other projects we are working on
We are currently working on a celebrity Look Alike app, that shows you the celebrity that looks most similar to you. We will launch it soon (Signup to our newsletter).
Furthermore, there are some updates on our roadmap for this attractiveness test.
Science of attractiveness
Research on attractiveness deals with the study of the attractiveness of human beings, specifically with regard to their physical features, their characteristics, their social position, their professional success, their material prosperity, their personality charisma. It cannot be assigned to any particular field, but is pursued in a variety of scientific disciplines, such as psychology, neuroscience, behavioral research, or even economics.
Do you save or publish my picturesThe images you upload are processed on our server, but we never save or publish them. We just keep the result for statistics without the photo.
Do you provide an API that I can use?There is no API yet, but we are considering creating one, as there seems to be demand. You can signup for our newsletter to get notified when we create an API. However we are currently quite busy with other projects, so don't expect anything in the near future.
Can I download the AI-model?No, but we are considering creating a standalone version and publish it to GitHub, as there seems to be interest in the research community. We will link it on this site, when we do it. Publishing the code of the current setup would not make much sense, as it relies on various tools in the Google Cloud.
If you like this tool, you can support us by sharing the site
The attractiveness test is free. No registration is needed, you just need to upload a picture. And of course, we don't save or publish the pictures you upload!
The results of this online tool are an indicator if you are hot or not. Try to follow our recommendations and please don't take to result too seriously if you don't like it.
Our hotness prediction model returns the result within seconds. If the site freezes for more than a minute, reload the page and try again.