Algorithm 2 Hierarchical STPS method
1:	i:=1; assume that an parameter net is given
2:	repeat
3:	Create icosahedral samples at level i for α’s and β’s
4:	if i=1 then
5:	Rotate the parameter net using each (αβ0), apply STPS, and keep the top K candidates that minimize the distortion costs, defined in Section 4.2.3
6:	else
7:	Keep only local icosahedral samples for α’s & β’s
8:	Rotate the parameter net of each top K candidate using each (αβ0), apply STPS, and select top K candidates
9:	i:=i+1
10:	until The best distortion costs do not improve
11:	Return the best result in top K list