[en] A comparison of two methods of scatter compensation in single photon emission computed tomography (SPECT) imaging is made on the basis of improvements in quantification. The methods, scatter-window subtraction and constrained deconvolution of an average point source response function (PSRF), are described; the theoretical basis of each method is also briefly assessed. Improvements in relative quantification offered by each method are measured by examining ratios of counts in hot cylinders to counts in a slightly radioactive background. The cylinder/background concentration ratio was varied by a factor of five; the sizes of the cylinders remained constant. Keeping the diameter of the cylinders constant allowed for an assessment of the effect of concentration differences, without the conflicting effect of variation in the size of the hot source. Results showed that while both scatter-window subtraction and constrained deconvolution offer quantification improvements in the final image, the method of prereconstruction deconvolution of a planar PSRF from each planar projection is substantially more successful than either scatter-window subtraction or other methods of implementing the deconvolution procedure