Figure 1From: Covariance tracking: architecture optimizations for embedded systemsCovariance descriptor computation. The image is first decomposed into an array of feature images (feature image tensor) applying the feature map F uv = ϕ (I,p uv ). Then the crossed-products of these features are computed; using these arrays, the tensor integral images A u ′ v ′ (i) and the second order integral images tensor B u ′ v ′ (i,j) are computed.Back to article page