Using K-means clustering algorithm built from scratch in Numpy to segment gray-scale images
Run the FasterKmeans.py for Grayscale Input Images.
Following is the image with original images encoded with 8 bits and a clustered image at 3 bits.
| Original 8-bit Grayscale image | Kmeans-based clustered image with 3-bit representation |
|---|---|
 |
 |
For colored images, run the kmeans(updated).py
Following is the clustered image at 8 different clusters.
| Original 24-bit Color image | Kmeans-based clustered image with roughly 5-bit representation |
|---|---|
 |
 |