I'm new to machine learning implementation and I'm trying to implement the image segmentation approach from the Blobworld paper. I have extracted the 6 feature vectors of the image that are described in the paper and I have them saved in matrix X which has 6 rows that are each a feature and number of pixels from the input image as columns.
I have managed to write the code that extracts the features. and now I need to cluster the feature vectors into gaussian clusters using EM:
X = double(computeBlobworldFeatureVectors(original_image));
X = X([1 2 3 6 5 4], :);
I am now trying to do (d) from the following image that is snapped from the paper :