Modern stroke research and treatment depend on the ability to accurately characterize and quantify diffusion volume. Current methods for segmentation and quantification of diffusion volumes are largely manual and time intensive. In this paper, we explored and validated a method of automatic segmentation of diffusion volumes. This algorithm was validated against known values from previous studies. We then studied the number of angles required to accurately predict diffusion volumes. We determined accurate volumes can be determined with as few as 3 directional vectors while accurate infarct mapping requires only 7 allowing for reduced scan time.