Seislet-based morphological component analysis using scale-dependent exponential shrinkage |

MCA considers the complete data to be the superposition of several morphologically distinct components: . For each component , MCA assumes there exists a transform which can sparsely represent component by its coefficients ( should be sparse), and can not do so for the others. Mathematically,

The above problem can be rewritten as

We prefer to rewrite Eq. (12) as

(13) |

Thus, optimizing with respect to leads to the analysis IST shaping as Eq. (9). At the

The final output of the above algorithm are the morphological components . The complete data can then be reconstructed via . This is the main principle of the so-called MCA-based inpainting algorithm (Elad et al., 2005).

Seislet-based morphological component analysis using scale-dependent exponential shrinkage |

2021-08-31