Clutter suppression in ultrasound: performance evaluation and review of low-rank and sparse matrix decomposition methods

Table 17 The most robust algorithms with the best performance

Group	Abbreviation	Algorithm name
RPCA	FPCP	Fast PCP
RPCA	L1F	L1 filtering
RPCA	DECOLOR	Contiguous outliers in the low-rank representation
RPCA	RegL1-ALM	Low-rank matrix approximation under robust L1-norm
RPCA	MoG-RPCA	Mixture of Gaussians RPCA
RPCA	Lag-SPCP-SPG	Lagrangian SPCP solved by spectral projected gradient
RPCA	Lag-SPCP-QN	Lagrangian SPCP solved by Quasi-Newton
RPCA	PRMF	Probabilistic robust matrix factorization
RPCA	GoDec	Go Decomposition
RPCA	SSGoDec	Semi-soft GoDec
RPCA	GreGoDec	Greedy semi-soft GoDec algorithm
MC	IALM-MC	Inexact ALM for matrix completion
MC	LMaFit	Low-rank matrix fitting
MC	LRGeomCG	Low-rank matrix completion by Riemannian optimization
LRR	ROSL	Robust orthonormal subspace learning
TTD	MAMR	Motion-assisted matrix restoration
NMF	PNMF	Probabilistic nonnegative matrix factorization
NMF	nmfLS2	Nonnegative matrix factorization with sparse matrix
NMF	Semi-NMF	Semi-nonnegative matrix factorization
NMF	Deep-Semi-NMF	Deep semi-nonnegative matrix factorization
TD	HoSVD	High-order singular value decomposition
TD	HoRPCA-S-NCX	HoRPCA with singleton model solved by ADAL (nonconvex)
TD	Tucker-ADAL	Tucker decomposition solved by ADAL
TD	Tucker-ALS	Tucker decomposition solved by ALS

ISSN: 1475-925X