Emojify

Inspiration

While exploring Microsoft's various cognitive API's we stumbled upon the uses of facial recognition. We realized that people had massive interest in social media such as SnapChat, which uses filters to augment people's faces. Thus, we got the idea to transfer peoples' faces in images to emoji characters. Building on this, we realized that we could connect this to another social media platform: Twitter. So we experimented with using Twitter's API to receive users' images and retweet the edited images back at the user.

What it does

Emojify essentially receives a tweets from Twitter users that contain the marker @EmojifyOfficial. Then, we took the images from the tweets and passed it to another process that identified human faces in the image. Then we logged the location of the faces and replaced them with appropriately sized emoji faces. However, importantly, our program can detect the emotions displayed by the people in the image and display the corresponding emoji image. After this, the edited image is tweeted back at the original tweeters.

How we built it

We built this using Microsoft's Facial Recognition API.

Challenges we ran into

Some of the challenges we ran into were researching the functionalities of the various APIs used such as the Twitter API, which was especially tedious. There was also the issue of effectively passing the edited image to be outputted from the facial recognition program to the Twitter bot.

Accomplishments that we're proud of

We're proud of actually successfully using the MIcrosoft API for something. It was interesting to get to learn how to use its built-in functionality to create an application of our own. Also proud of the fact that much of our team is comprised of first time hackers. We are immensely satisfied with being able to build something that works and can interact with others via social media.

What we learned

We learned a lot of stuff about harnessing the power APIs as well as using Node.js. It was awesome being able to learn more about how we could connect different technologies together.

What's next for Emojify

We want to be able to learn how to train our own models to recolonize faces instead of just using apis. We think we could really create something amazing in the future with the skills we learn from building Emojify!

Built With

• computer-vision
• emotion-recognision
• facial-recognision
• html5
• javascript
• microsoft-apis
• node.js
• twitter