Scale Selection for Anisotropic Scale-Space: Application to Volumetric Tumor Characterization

A unified approach for treating the scale selection problem in the anisotropic scale-space is proposed. The anisotropic scale-space is a generalization of the classical isotropic Gaussian scale-space by considering the Gaussian kernel with a fully parameterized analysis scale (bandwidth) matrix. The "maximum-over-scales" and the "moststable-over-scales" criteria are constructed by employing the "L-normalized scale-space derivatives", i.e., responsenormalized derivatives in the anisotropic scale-space. This extension allows us to directly analyze the anisotropic (ellipsoidal) shape of local structures. The main conclusions are (i) the norm of the - and L-normalized anisotropic scale-space derivatives with a constant =1/2 are maximized regardless of the signal's dimension iff the analysis scale matrix is equal to the signal's covariance and (ii) the most-stable-over-scales criterion with the isotropic scale-space outperforms the maximum-over-scale...

• Type: Conference
• Year: 2004
• Where: CVPR
• Authors: Kazunori Okada, Dorin Comaniciu, Arun Krishnan
• Keywords: Analysis Scale Matrix, Anisotropic Scale-space Derivatives, Anisotropic Spreads, Computer Vision, CVPR 2004, Isotropic Gaussian Scale-space, L-normalized Scale-space Derivatives