Citation Context

...pt ) ∧ ( d > threshold ⇒ reject ) . (1) 3 The Eigenface Method In this section we give a brief explanation of the eigenface method of face recognition, while referring the reader to Turk and Pentland =-=[2, 3]-=- for more detailed explanations. We compute the covariance matrix C, of facial images from a set of M (60) training images: {Γ1, Γ2, Γ3, Ö ΓM}, C = M 1 = M ∑ n =1 AA T Φ n Φ T n A = [ Φ1 Φ2Φ3... ΦM ] ...

Citation Context

...an be applied to whole face-key vectors, rather than individual vector elements, providing a global discriminant value d for the entire face space, d = c ∑ c ∑ i= 1 ω − ω ∑ 1 Φi i= 1 ω∈Φi i ω − ω i . =-=(6)-=- where ω is the face-key of some single or combined face space, to which we apply the recognition distance metric (equation 4) to the averageω and class average ω . Applying equation 6 to each fisherf...

Citation Context

...Two broad approaches to the segmentation and tracking of moving objects are motionbased and model-based. Both methods have their relative strengths and weaknesses. Current motion-based approaches [26]=-=[3]-=- at segmenting moving faces from an image sequence have used pixel based, simple change detection procedures based on difference images. These techniques may run into difficulties when multiple moving...