I'm trying to implement the Eigenfaces algorithm and was able to check if a face matches by checking the euclidean distance as shown below. I would like to improve the code using Mahalanobis Distance instead.
similarity_score = arrayfun(@(n) 1 / (1 + norm(features(:,n) - feature_vec)), 1:num_images);
where features is a table containing a feature in each column and feature_vec is the feature vector we're checking.
