Classification

• Classification with two inputs (x, y) and one output (color label)
• Classification with four inputs (x & y of fingers relative to palm base) and one output (”scissors”, “stone”, “paper”), based on HandPose
• Classification with an image as an input (64x64 pixels x4 color channels) and one output (label)

Regression

• Regression with two inputs (x, y) and three outputs (r, g, b)

See also:

• Daniel Shiffman: Nature of Code, 2nd Edition, Chapter 10: Neural Networks