Week 12

It finally happened!! I built my first basic classifier and IT WORKS!!

After numerous hours of debugging, I finally managed to implement a basic classifier that is able to differentiate between the discrete emotions neutral, happy and sad in a very expressive context using only the Muse headband. One aspect that still needs some revision though is the feature selection, which is non-existent at the moment because of some residual bugs…

The next thing on my agenda was the implementation plan for the feedback mechanism discussed last week. After Karsten sent me on the right path, I found an interesting spinning wheel to represent a small set of discrete emotions:



My plan is then to overlay the image with an arrow and let that arrow point to the emotion that was detected automatically. After that, the user should be able to manually adjust the arrow to point to the emotion that he or she is actually feeling. In order to help me get started on the implementation, I found a few Android sample projects and tutorials: Project 1Project 2Project 3.

To conclude this post, I thought it would be interesting to provide a link to a recently published article on DARPA’s vision of the future. Let’s make this future happen and start recognizing some emotions!!


2 thoughts on “Week 12

  1. Hey Jeroen!

    Nice idea, this spinning wheel.
    Are you keeping these emotions, or are you planning to change them?
    I think “trust” and “surprise” are maybe not the most appropriate emotions..
    Surprise is in my opinion a very short emotion. You feel surprised for just a moment, and then you feel the emotion the surprise triggered, like happiness when you get a car for your birthday, of disgust when you see something horrible in your parents bedroom.

    And what about excitement, love, neutral? Maybe even tiredness?

    Maybe you can shrink the arrow for a more neutral happiness.
    I don’t know if you can detect such differences.

    Tschüss !


    1. Yes, you couldn’t be more right, Eline! Of course, every software engineer tries to keep his or her code as adaptable as possible, so adding or removing some emotions from the wheel should not be a problem.
      About that shrinking arrow: you might be on to something there! I will certainly take your idea into consideration while implementing the emotion wheel. Nevertheless, I must say that I was planning on following the discrete approach instead of the dimensional one due to the inherent measurement and classification difficulties of the latter (see my first posts on the literature study).


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s