Driver head pose estimation using efficient descriptor fusion

EURASIP Journal on Image and Video Processing

Table 1 Overview of the most relevant literature approaches

Reference	Year	Type	Methods	R1	R2	R3
Our ^(+,0,4)	2015	Cl	Descriptor fusion + SVM	√	√	√
[17] ⁽¹⁾	2012	GM	Face symmetry	√	√	×
[12] ⁽⁰⁾	2012	GM	3D geometry	√	√	√
[18] ^(+,4)	2013	GM	Golden ratio	√	√	×
[19] ^(2,3)	2010	FM	Face model	√	√	√
[21] ^(+,4)	2013	AT	SF + LPF	√	√	√
[31] ^(+,4)	2012	Rg	HOG + PLS Rg	√	√	√
[24] ^(+,4)	2012	Rg	SIFT + RP	√	√	√
[23] ^(+,4)	2013	Cl	Multi-scale SF + SVM	√	√	√
[27] ^(+,4)	2014	Cl	Joint detection and estimation + SVM	√	√	√
[28] ^(+,4)	2014	Cl	Gabor + covariance + learning	√	×	√
[26] ^(∗,4)	2007	Cl	Associative memory	√	√	√
[32] ^(∗,4)	2008	Cl + Rg	SVM + SVR	√	×	√
[35] ^(∗,4)	2007	DR + Cl	LDA + linear learning	√	√	√
[6] ⁽⁰⁾	2008	Tr	Tracking using particle filters	√	√	√

Best result approaches (plus sign), most used references for benchmarks (asterisk). Databases: “0”: Own; “1”: FacePix [48]; “2”: BU [49]; “3”: MIT [50]; “4”: Pointing’04 [34]
GM geometric model, FM flexible model, AT appearance template, Cl classification, Rg regression, Tr tracking, DR dimensionality reduction