Week 2: February 8th, 2023

Working with: KNN Image Classifier / teachable machines - image project and pose project

Ideas

1. Image Classification

Increase color of square or background by showing different colored object (collection of red objects to make background more red, blue for more blue etc.

If switching to HSB color mode, can include more than just red, green, and blue)

2. Pose Classification

yes/ no gestures and making decisions with yes/no questions asked nested conditional statements like the choose your own adventure shows. But can be for like baking/following recipe or customizing home screen/canvas If I’m worried about it going too fast, I can add another class where if shown, just do nothing. Then it’ll detect like you sitting there and not do anything vs its looking for one of the categories and doing something you don’t want it to do

I went with idea 2 because I previously worked a little with image classification already (ICM Week 9: Body) and wanted to try something new.

For 3 questions, it expanded to 4 sets of “if yes / if no” statements within a decision tree.

I am trying to make a background/wallpaper using the questions:

1. Do you want to switch to dark mode? yes/no
2. Do you like squares more than circles? yes/no
3. Do you want to see it move? yes/no

I drafted out the decision tree prior to running everything. This is how the conditional statements looked in code, p5 Sketch: Background Making Code