I am working on a human identification by gait analysis project. So far, I have managed to extract the Gait Energy Image(GEI) of a silhouette. I am stuck on finding a way to move forward with my project. I am thinking about applying PCA on the GEI of each silhouette and then using kNN with k = 1 for classification. Does PCA work with GEI? And what would be a better method instead of kNN for this particular project?