Trained 25 face images (each having a dimension of 425 by 425) by implementing the Eigenface Algorithm and performed the following steps:
- Calculation of the mean face.
- Performed Principal Component Analysis (PCA) to get the Eigenvectors of the covariance matrix of the training images.
- Reconstructed the training faces by selecting top k = 2, 5 and 15 Eigenfaces (eigenvectors which correspond to the largest eigenvalues).
- Tested the algorithm on 32 images. Projected each test image on top k = 2, 5 and 15 Eigenvectors and classified it as a face or nonface. Recognized the face images by using Eucledian distance to find the closest training image.
- Plotted a graph representing percent classification error rate as a function of the number of Eigenvectors(k)