In this part, you will use unsupervised learning to reduce the number of color points in an image. This application is also referred to as color quantization. You are free to use either k-means or EM algorithm for the unsupervised learning. You are also free to use any suitable library or package, but you have to clearly specify the name and installation instructions in the README file. Below are some resources for learning more about color quantization:
- https://en.wikipedia.org/wiki/Color_quantization
- https://lmcaraig.com/color-quantization-using-k-means/
- https://appliedmachinelearning.blog/2017/03/08/image-compression-using-k-means-clustering/
Input: You are given a set of 5 images that you can download from: http://www.utdallas.edu/~axn112530/cs6375/unsupervised/images You can select any 3 out of these and run the color quantization algorithm on them.
Parameters: It is up to you to select the best value of the number of clusters and any other parameters for the algorithm. You should try at least 3 of k values for each image.
Output: You should include the output of the 3 images i.e. quantized images for any one value of k. Create a folder called "quantizedImages" and place the images there.
How to run: Please include instructions on how to run your code in the report file. Do not use hard-coded paths or parameters.