Hi, This is fine as long as you just want to compare image1 vs image2 with *no* transformation. Generally, I expect you'd want to use -init to specify a transformation. For example, do a registration first (no schedule file) and save the (transformation) matrix, then use this as the -init for this flirt call with measurecost1.sch. Also, you don't need to specify searchcost with the measurecost1.sch. The refweight or inweight masks are always taken into account with any cost function calculations - either in a registration or in the measurecost1.sch. All the best, Mark On 9 Dec 2008, at 22:04, siamak wrote: > Hi, > > Actually, I am using a command like this: > > flirt -in image1.nii -ref image2.nii -cost cost1 -searchcost cost1 - > schedule > measurecost1.sch -refweight mask > > Is the result of similarity measure is representing the alignment in > mask volume? > > Siamak >