Goal: Characteristics of vortices within intracranial aneurysmal flow patterns have been associated with increased risk of rupture. The classifications of these vortex characteristics are commonly based upon qualitative scores, and are therefore subjective to user interpretation. We present a quantitative method for automatic time-resolved characterization of 3D flow patterns and vortex detection within aneurysms. Methods: Our approach is based upon the combination of kernel deconvolution and Jacobian analysis of the velocity field. The deconvolution approach is accurate in detecting vortex centers but cannot discriminate between vortices and high-shear regions. Therefore, this approach is combined with analysis of the Jacobian of the velocity field. Scale-space theory is used to evaluate aneurysmal flow velocity fields at various scales. Results: The proposed algorithm is applied to computational fluid dynamics and time-resolved 3D phase-contrast magnetic resonance imaging of aneurysmal flow. Conclusion: Results show that the proposed algorithm efficiently detects, visualizes, and quantifies vortices in intracranial aneurysmal velocity patterns at multiple scales and follows the temporal evolution of these patterns. Significance: Quantitative analysis performed with this method has the potential to reduce interobserver variability in aneurysm classification.