Human faces, e.g. facial parts, sketches or expressions

eigenface eigenfaces face representation using a principal component analysis in a high-dimensional space created from images of faces. The eigenvectors of the representation are derived from the covariance matrix of the probability distribution computed in this high-dimensional vector space. fisherface fisherfaces linear discriminant analysis (LDA) applied in a multi-dimensional representation space created from a set of face images, resulting in a set of basis vectors defining that space frontal face recognition face images are generally obtained by placing a camera in front of the subject who is asked to look at the camera while the picture is taken illumination-invariant recognition recognition insensitive to changes in lighting conditions multiview face recognition employs a gallery of images of every face at various poses to cover multiple views for each face pose-invariant recognition recognition insensitive to changes in pose