#duoproject

After a talk with one of the teachers, I have decided to give up on emotion detection with a camera. I spent a lot of hours on it and I’m sad to give up on it, but it’s for the better to drop this project that was taking too much time without any real result.

Sometimes, after hard work, things still don’t work out. On paper everything in this python application should work, but instead of getting a live video feed, I’m getting static.  There are no errors in the code (if my webcam doesn’t work, there should be an error) so there might be something wrong with the webcam settings on my computer.

When trying to do mahcine learning, make sure you have enough memory in your pc, or don’t do it on your own machine.

Sometimes things just don’t work out the way you want them to... There were problems with a library I needed to make the emotion detection work on my Raspberry Pi. Likely, the cause of this is the creators of the library not updating it.

I once heard a classmate say “The Raspberry Pi does everything a computer does, just worse.” and it’s safe to say that I think he’s right.

Therefore we decided to develop our project on a computer with Windows 10 first, to get the proof of concept out. After that it’s possible to look at transporting it to a Raspberry Pi so it can actually fit into a robot.

This little guy will have to help use with our face/emotion detection journey. I found a promising tutorial using Python, for an emotion recognition program.

So what should the robot be able to do? At the very least we want the robot to be able to see the emotion on the user’s face, and act accordingly in a positive way to motivate the user.

Sketching definitely is an important method to draw out your concept, but it isn’t really tangible.

This is where paper prototyping comes in handy. This is a good method to see how your product would look in the real world (in a lo-fi way). You can see what works and what doesn’t work with the physical product. For instance, I immediately noticed that the dimensions of the robot we had in mind were way too big.