Science 9

 

My science fair project is an innovation. My innovation is going to be an algorithm (created with machine learning), which will attempt to identify emotions in pictures, and videos. My goal for the fair is to have trained the algorithm enough, that it will have a 60-80% success rate. Above is a screenshot of a sample dataset, and a simple input reader in the case that I choose to use a GUI. If I need to contact an expert, I can contact one of the many experts online, or the IT teacher. I could contact one of the writers of an article or paper (assuming they would reply), or as I mentioned earlier, I could contact the IT teacher in the school, who may be able to provide me with guidance. I have quite a bit of background knowledge on the subject, because I have been following machine learning for a while now. Some of the articles I have read on the subject which helped my research are:

InfoQC – EmoPy

Emerj – What is machine learning?

DigitalOcean – Intro to machine learning

As well as this Coursera course

After spring break:

I have now completed the program. Every second, it will take a picture, identify any faces in that picture, and then attempt to guess the emotion based on three tests. One decides between anger and happiness, one decides between calm, anger, fear, and surprise, and the last decides between sadness, disgust, and surprise. Here is a screenshot of my work.

 

Post-Fair Reflection:

1. Yes, I was able to add to my Edublog to reflect on science fair.

2. I did not work with a partner.

3.  My solution answered my original question, which was “Can an AI identify emotions?” The answer was yes, as my algorithm can identify emotions with fairly high accuracy.

4. I ended up using technology to create a solution. Specifically, I used python to code a solution. As I have mentioned previously, it went well, as my algorithm is fairly accurate.

5. I conducted research on how machine learning works, how face recognition works, how human emotions appear, and how to feed machine learning algorithms data.

6. While I did not directly communicate with any experts, I read some papers by experts in the field, as well as courses by said experts.

Skip to toolbar