Practice: Duckie Hat 🦆🎩#

Download starter code

Write a function called duckie_hat that takes an RGB image of a duck (a numpy.array with shape (Height, Width, Color Channels) ) and returns a new image with the same pixels but modified so the duck is now wearing a fancy green top-hat.

Our algorithm won’t do anything fancy like detect where the duck is. Instead, we will just hard-code where the hat should be.

You should make a hat by setting all the pixels in the following regions to be fully green and the other channels off. The locations below are in terms of (0, 0) being the top left and the first dimension going down the image and the second dimension going to the right.

  • The main-hat should be a rectangle with the top-left corner at (20, 85) and the bottom-right corner at (75, 135) .

  • The hat brim should be a rectangle with the top-left corner at (75, 60) and the bottom-right corner at (90, 160) .

Warning

Clarification: For simplicity, we intended for the specification of the bottom-right corner to be exclusive but failed to state that explicitly. This means for the main hat, the hat should go up to, but not including the point (75, 135) .

For the duck image from the reading, your output image would look like the following:

TODO

Implementation Details#

  • Your solution should use no loops or comprehensions to solve this problem.

  • You should not modify the input image. Instead, you should make a copy of it with the .copy() method provided by numpy arrays.

Note

Credit goes to Mitchell Estberg for coming up with this idea 🦆🎩