S ← Read Image

for every x ∈ domain(S) do

Γ ← [x − w, x + w] × [y − w, y + w] × [z − w, z + w]

Compute η(x)

Compute optimum λb, for the given η

Lmax ← −1

for i = 1 to N do

Compute optimum λ0, λ1 for the given η

if ${\lambda }_1-{\lambda }_0<{\lambda }_T^l$ then

Continue

else

Compute the α-ranked likelihood ratio L

if Lmax ← L then

posmax ← i

Lmax ← L

λC = λ1 − λ0

end if

end if

end for

if Lmax > 0 then

if ${\lambda }_C\ge {\lambda }_T^h$ then

Mark(x) ← true

end if

Normal(x) ← posmax

Likelihood(x) ← Lmax

contrast(x) ← λC

end if

end for