Academic project. Part of the Msc in Statistics and Operational Research at UPC-UB
Eigenfaces is the given name for the eigenvectors when applyig a principal component analsyis to a collection of images. From a picture dataset of all the members of the parliament of 2008, we have performed an unsupervised learning technique to solve a human face recognition problem. The idea of the project was to detect the main facial features of all deputies, to later use a non-linear classifier to assess the gender from just the image of the representative.
The work aimed to demonstrate the theoritical knowledge of multidimensional scaling, clustering, and classification algorithm, and their implementation through the python scikit-learn package.