engleski

Blind decomposition of low-dimensional multi-spectral image by sparse component analysis

Multilayer hierarchical alternating least square nonnegative matrix factorization approach has been applied to blind decomposition of low-dimensional multi-spectral image. Performance of the algorithm is invariant with respect to statistical (in)dependence between materials present in the image that is an assumption upon which many existing blind source separation methods depend. The proposed method performs blind decomposition exploiting spectral diversity and spatial sparsity between the materials present in the image. Unlike many existing blind source separation methods the method is capable to estimate the unknown number of materials present in the image. This number can be less than, equal to or greater than the number of spectral bands. Performance of the method is evaluated on underdetermined blind source separation problems associated with blind decompositions of experimental red-green-blue images composed of four materials. The proposed algorithm showed best performance in comparison with methods based on -norm minimization: linear programming and interior-point methods. In addition to tumor demarcation problem demonstrated in the paper, other areas that can also benefit from proposed method are cell and chemical imaging.

Multi-spectral imaging; Chemical imaging; Cell imaging; Sparse component analysis; Nonnegative matrix factorization.

nije evidentirano