Table 2.

Mapping of various popular recognition algorithms to the canonical architecture of Figure 5.

Method	Stage 1		Stage 2
	Units	Saliency	Templates	Units	Assignment	Pooling
HMAX	Filter responses	–	Random	LU	Soft	Max
MPE	Filter responses	–	GMM	CPU	Soft	Sum
NBNN	SIFT	Bottom-up	Training set	CPU	Hard	Sum
SPMK	SIFT	Bottom-up	Codebook	PU	Hard	Sum
HGMM	SIFT	Bottom-up	GMM	CPU	soft	sum
Sparse SPMK	SIFT	Bottom-up	Sparse dictionary	PP	Soft	Max
LSN	SL	Top-down	–
HDSN	DS	Top-down	Random	DS	–	Sum

HMAX: (Serre et al., 2007), MPE: (Carneiro et al., 2007), NBNN: (Boiman et al., 2008), SPMK: (Lazebnik et al., 2006), HGMM: (Zhou et al., 2009), sparse SPMK: (Yang et al., 2009), LSN: (Elazary and Itti, 2010).