me_llt2d - 2D Moreau envelope, LLT algorithm
Compute numerically the discrete Moreau envelope of a set of spatial points (Xr(i1),Xc(i2),f(i1,i2)) at slopes (Sr(j1),Sc(j2)), i.e.
2 2 M(j1,j2) = min [ f(i1,i2) + (Sr(j1) - Xr(i1)) + (Sc(j2) - Xc(i2)) ]. i1,i2It reduces computation to one dimension, and computes the Legendre conjugate through the formula
2 2 M(j) = s(j) - 2 g*(j) with g*(j) = max [ s(j) * x(i) - 1/2 * (x(i) + f(i)) ] ithereby resulting in a theta(n*m + m1*m2) linear-time algorithm.
function f=f(lambda,x),f=lambda * x.^2,endfunction function g=g(lambda1,lambda2,x,y),g=f(lambda1,x)+f(lambda2,y),endfunction lambda1=1;lambda2=2; x1=(-10:10)';x2=(-5:5)'; [X, Y]=ndgrid(x1,x2);F=g(lambda1,lambda2,X,Y); s1=(-4:4)';s2=(-5:6)'; Xr=x1;Xc=x2;Sr=s1;Sc=s2; desired=me_llt2d(x1,x2,F,s1,s2); //1d computation for separable function Ms1=me_direct(x1,f(lambda1,x1),s1); Ms2=me_direct(x2,f(lambda2,x2),s2); t1 = Ms1 * ones(1,size(Ms2,1)); t2 = ones(size(Ms1,1),1) * Ms2'; correct=t1+t2; b = and(correct == desired);
me_brute2d, me_direct2d, me_nep2d, me_pe2d, me_llt,
Yves Lucet, University of British Columbia, BC, Canada