The article presents a technique for computing 2D component velocity from image sequences using a family of spatiotemporal velocity-tuned linear filters. The component velocity is derived from the local first-order behavior of surfaces of constant phase, which is justified from the perspectives of 2D image translation and deviations from translation in perspective projections of 3D scenes. The method is linear, efficient, and suitable for parallel processing, with robustness to noise and the ability to estimate multiple velocities within a single neighborhood. The authors focus on component velocity (normal velocity) to achieve more accurate estimates of motion within smaller apertures, enhancing spatial resolution of velocity fields. The approach contrasts with other methods like gradient-based, correlation-based, and contour-based techniques, emphasizing the use of phase information for better velocity resolution, subpixel accuracy, and robustness to smooth contrast changes and affine deformations. Experimental results with realistic image sequences demonstrate the technique's accuracy and robustness, including cases with significant perspective deformation.The article presents a technique for computing 2D component velocity from image sequences using a family of spatiotemporal velocity-tuned linear filters. The component velocity is derived from the local first-order behavior of surfaces of constant phase, which is justified from the perspectives of 2D image translation and deviations from translation in perspective projections of 3D scenes. The method is linear, efficient, and suitable for parallel processing, with robustness to noise and the ability to estimate multiple velocities within a single neighborhood. The authors focus on component velocity (normal velocity) to achieve more accurate estimates of motion within smaller apertures, enhancing spatial resolution of velocity fields. The approach contrasts with other methods like gradient-based, correlation-based, and contour-based techniques, emphasizing the use of phase information for better velocity resolution, subpixel accuracy, and robustness to smooth contrast changes and affine deformations. Experimental results with realistic image sequences demonstrate the technique's accuracy and robustness, including cases with significant perspective deformation.