Face verification using tensor representation from video

We present a new face verification method based on tensor presentation from video. For this purpose, we use Tensor-based Multilinear Principal Component Analysis (TMPCA) to verify simultaneously the successive frames with face images. This approach extracts feature directly from third-order tensor representation rather than the well-known vector or second-order tensor representation. The experimental results on face image database with distance-varying reveal that the presented method shows a good verification performance.