A proof-of-concept to use K-Means Algorithm for Image Vectorization.
- Convert the Image into Frequency domain
- Separte beackground and foreground using high pass, band pass and low pass filters
- Find a suitable number of number of centroids in K-Means
- Apply K-Means to both the background and foreground
- Recreate the Image
- Apply Image Sharpening to make edges more prominent