Cloudinary Blog
Build a Facial Emotion Recognition Based Video Suggestion App

Developers are always looking for new and creative ways to deliver content that resonates with the way users feel. Often using the latest technical innovations the market has to offer such as Artificial Intelligence (AI) and Machine Learning (ML). What better way to demonstrate innovative uses of these technology in a consumer market than capturing expressions from your users and then serving content based on that expression!

In this article we are going to build an app that suggests videos to users based on their facial expressions (i.e. emotion). To do this, we will use Cloudinary’s Advanced Facial Attributes Detection Add-on and the Cloudinary Video Player.


Sign up for Cloudinary free today!


Set Up Cloudinary

Cloudinary makes adding image and video optimizations to applications a breeze. Head over to the Solutions Page to learn more about the features offered.

First, sign up for Cloudinary. Once you have created an account, you are able to use the “Advanced Facial Attributes Detection” add-on.

You can add the “Advanced Facial Attributes Detection” to your Cloudinary account here and follow the on-screen instructions.

Advanced Facial Attributes Detection Add-on

Make sure to note your API_KEY , API_SECRET and CLOUD_NAME from your developer console. This information is needed when integrating with your application.

Upload Images and Retrieve Facial Emotion Recognition from Images

Using Cloudinary to obtain the emotion from a facial image requires the Advanced Facial Attributes Detection Add-on when uploading the image. We add the adv_face as a tag to let Cloudinary know we are going to use the Advanced Facial Attributes Detection add-on to register the emotions from the detected users face.

When the image is successfully uploaded, emotions detected from the face and their confidence values are returned as part of the result. The emotion with the highest confidence is then selected.

JavaScript Sample:

Copy to clipboard
cloudinary.uploader.upload( req.body.image , function(result) {
  const emotions = result.info.detection.adv_face.data[0].attributes.emotion;
  let arr = JSON.parse(JSON.stringify(emotions));
  let visible_emotion = getMaxKey(arr); // this function gets the emotion with the highest confidence
  return res.json({
    status: true,
    emotion: visible_emotion
  })
},{ detection: "adv_face" });

You can then return the detected emotion to the frontend of your application.

Creating Video Playlists Based on emotion

Once the emotion is obtained on the front end, we want to display a set of videos to the user that correspond to the selected emotion. The Cloudinary Video Player makes creating playlists simple. You can create a video player and then populate it with videos from your media library that are tagged with the highest confidence emotion, such as ‘happiness’.

JavaScript Sample:

Copy to clipboard
let emotion = "happiness";
let cld = cloudinary.Cloudinary.new({ cloud_name: CLOUD_NAME, secure: true});b
// initialize video player
let demoplayer = cld.videoPlayer('elementID');
//  create playlist based on emotion
demoplayer.playlistByTag( emotion, { sourceParams: {angle:0}, autoAdvance: 0, repeat: true, presentUpcoming: 5});

Putting it all together, we have the sample of how it all works below:

Video Suggestion at Work

Conclusion

In this post, we have shown how to use the Advanced Facial Attributes Detection add-on from Cloudinary to create an application that enables us to serve videos to our users based on how they are feeling at the moment. We also leveraged the Cloudinary Video Player to serve the videos in a playlist format. You can learn more about using the video player and feel free to check out the github repository for the full source code. If you are up for the challenge, clone and submit your own Video Suggestion App below!


Further Reading on Video Manipulation

Recent Blog Posts

The Benefits of Headless DAMs

Headless is not a buzzword anymore. In fact, the concept of headless architecture is gaining momentum due to the flexibility it offers for composing new experiences and for tackling the undue complexity of an ever-evolving technology stack. That’s because while the evolution of the martech landscape has enabled disruptive, digital innovations, the approach of buying point solutions for solving specific challenges can expose companies to the complicated nature of new technologies, systems, and platforms.

Read more

Building Display Ads With Transparent Video

By Afzaal Ahmad Zeeshan
Build Web Ads With Transparent Video to Attract User Engagement

Billions of views on the Internet every day drive one of the biggest industries on the planet: advertising. The sheer size of that market and the competitive nature of vying for consumer attention results in a constant need for innovation. Readers are jaded, and display ads are blind spots.

Read more
How Cloudinary's Media Optimizer Helps E-Businesses Deliver Superior Web Performance

As a technology company, Cloudinary owes its success to its ability to build solutions that address the most critical challenges you, our customers, face. The companies we serve run the gamut of digital businesses—retailers and direct-to-consumer brands, media and entertainment, travel and hospitality—which, coincidentally, all care about the same things.

Read more
How to Auto-Tag Video With Markers on Cloudinary for Easy Navigation

A picture is worth a thousand words, and that also holds true for video, one minute of which, according to Dr. James McQuivey of Forrester Research, is worth 1.8 million words. That's why online stores rely on rich media to promote products and sales. Images and videos impart a real sense of involvement with a purchase—a car, a vacation getaway, an apartment rental—setting your business apart from the competition.

Read more