Your Emotions Are Only Emojis

by Cassie Tarakajian

I successfully completed week one of my one-project-a-week challenge. Overall, I must say I'm feeling fairly disappointed. Maybe you'll understand why when you check out my project over at It doesn't work totally how I wanted it to, but I learned some things and made a thing and that's all that matters, right?

The Project

I love emojis. However, when I really think about it, I find it terrifying that our social interactions can be watered down to them. You can send your significant other a kissing face rather than say you love him or her, which feels less intense. Our emotions now sit in a discrete set of boxes, chosen by this guy. This can go even futher: if we have a quantized set of emotions, wouldn't a computer be able to figure out how you're feeling? Would then computers know more about us than we do?

This is something explored in many dystopian novels, including Do Androids Dream of Electric Sheep?, in which the protagonist Rick has total control over his emotions by dialing in the mood he wants into a mood organ. Or maybe moods will be maintained with a regimen of presription drugs, like in the film THX 1138. Humans make mistakes when experiencing intense feelings of stress, anxiety, or depression, so why not let a computer have control over your emotions? We should all always be happy, right?

(This does not my reflect my actual feelings on this topic, I just want to play devil's advocate for a bit.)

The Tools

To do the facial recognition and emotion detection I used a JavaScript library I found called clmtrackr. It works reasonably well, just as long as you don't move your head too much. It also doesn't work very well if you're wearing glasses.

This library works by using computer vision to match a generalized model of a face to your face.

Man with a face

cmltrackr also includes a basic emotion detection library, which determines emotions based on the positions of points within the facial model.

The Results

clmtrackr is easy to use and an awesome start for open source facial recognition and emotion detection. However, I was disappointed by how long it takes to find your face and how it loses focus of your face if you make sudden movements. The range of emotions it can detect is happy, sad, surprised, and angry, and even that matching is not so accurate. But some projects are kind of failures, and that's okay.

Cassie Tarakajian

hardware nerd, software nerd, web nerd, nerd nerd.

Read more from this author