EE 5450 Module 3 Homework 0
In this homework, you will practice using the OpenCV image processing
tools to find a Pokemon (either a Psyduck or a Pikachu). You'll notice two
Python files: segmenter.py and test_segmenter.py. The former contains
the class definitions and the latter contains sample tests. There are four
supplied test images in the samples/ directory, of which are all used for
testing. Note that you would normally have a training and testing data set,
but we are just practicing the image processing tools here and will explore
machine learning over the next lessons and assignment.
To get started:
-
I have used an interface-based design for this particular assignment.
SegmenterInterfaceis a class that requires its derived classes (meaning class that implement the interface) to implement its methods. The gist of each method is written as docstrings. Because the base interface has docstrings, you can just copy/paste the docstrings over to your implementations of those functions. -
Implement the
PikachuSegmenterclass first. This is simpler because the test images contains only one Pikachu. I recommend to tmplement the functions in order (feel free to implement any helper functions needed):a.
enhance_image()to perform CLAHE on the each channel.b.
threshold_enhanced_image()to get a noisy thresholded image of the object(s) you are looking for.c.
clean_thresholded_image()to clean up your noisy thresholded imaged.
get_combined_thresholded_image()to reduce the four-channel binary image to a single binary imagee.
get_bounding_boxes()to call all of these functions and spit out the bounding boxes (list of tuple of: x_left, y_top, width, height) that contain the object desired. -
Once you have the
PikachuSegmenterworking, try your hand at thePsyduckSegmenter.
Fair Use Notice
Please don't sue me, Pokemon Company International! I think this is considered fair use since it's for educational teaching purposes.