Matrix Methods for Geometric Data Analysis and Pattern Recognition Cats and Dogs Classification Project

In this project we were given 80 images of cats and 80 images of dogs which was to be used as a training set to classify another set of 38 images containing cats and dogs. In this project four different methods were used for classification. The first method uses the Average and Laplacian filter followed by Linear Discriminant Analysis. The second method uses an k-nearest neighbors search with an adaptive metric. The third method uses the Voronoi Diagram. Finally, the last method used the SVM (Support Vector Machine) method. The classification rates are 97.37%, 94.73%, 89.47%, and 86.84%, respectively.