Skip to Main Content
This paper describes example-based face detection from images involving complex background using ICA-based feature extraction and RDF recognizer. One of the difficult problems for example-based face detection is to estimate the lower-dimensional subspaces without reducing the discriminative features. For face recognition, ICA is a better tool to represent face images as a few components as possible. We used RBF neural network to distinguish faces and non-faces and evaluated its performance as discriminative properties using ICA.