Face recognition under non-uniform motion blur, illumination and pose variations

The current face recognition algorithms are unreliable when the pose, illuminating conditions of the test face is not the same as the stored face. In order to overcome the drawbacks of the current techniques in identifying the individual face in different atmospheric conditions such as blur, illumination and the pose, our paper proposes an approach which can identify the face irrespective of the variations. A pose contingent linear transformation of the identity vector in the presence of noise is implemented and the measured vector is generated. The blur face is modeled as a convex combination of geometrically transformed instances of the focused gallery face, and the set of all images obtained by non-uniform blur produces a convex set. The proposed method uses a feature extraction approach using CCA to extract the from the test face. The extracted features are compared with the features available in database to identify the person face.