How to impose a global parameter smoothness constraint on multiple task regression


New Member

I have a series of lets say 5 images. At each point of the image stack, I am fitting three parameters, {a0, a1, a2} using the 5 values from the image stack at that point.

So for a 100 x 100 pixel image I will get 10,000 different sets of parameters, {a0,a1,a2}. From the physics of the problem, a0 represents the average intensity value of the image at each pixel point. The average value map formed from the a0 values at each point should be smooth. I would like to impose a smoothness constraint on the a0 values while fitting all of the parameters at each point. Is this possible? Can someone point me in the right direction?