function out = max_depth_of_field(ims)
   
    dims = ndims(ims);
    
    %gradients are black and white, so they will loose a dimension
    %gradients = xy_gradient(ims(:,:,:,1));
    gradients = o_grad(ims(:,:,:,1));
    %gradients = contrast_map(ims(:,:,:,1),12);
    
    size_ims = size(ims);
    
    
    for i  = 2:size_ims(dims)
        
       %the + .075 * i is to bias towards images in the back 
       %gradients = cat(dims, gradients, xy_gradient(ims(:,:,:,i)));
       gradients =  cat(dims, gradients, o_grad(ims(:,:,:,i)));
       %gradients = cat(dims, gradients, contrast_map(ims(:,:,:,i),12));
    end

    max_grads = max(gradients,[],dims);
    
    a_image_red = zeros(size_ims(1:2));
    a_image_green = zeros(size_ims(1:2));
    a_image_blue = zeros(size_ims(1:2));
    
    for i = 2:size_ims(dims)
       a_image_red   =  a_image_red + ims(:,:,1,i) .*  ((max_grads == gradients(:,:,i)) & (a_image_red == 0));
       a_image_green = a_image_green+ ims(:,:,2,i) .*  ((max_grads == gradients(:,:,i)) & (a_image_green == 0));
       a_image_blue  = a_image_blue + ims(:,:,3,i) .*  (( max_grads == gradients(:,:,i) ) & (a_image_blue == 0));
    end
    
    out = cat(3, a_image_red, a_image_green, a_image_blue);