Right now the _sino_indicator() function does not use the weight values. This could be a problem if there are large erroneous values in the sinogram with weights of 0.

Should we pass the weights to the _sino_indicator() function?