The growth cone formed at the tip of the growing neurite is a highly motile structure that mediates its extension and steering. Although the growth cone motility is closely related to the mechanism of neurite growth, it has not been precisely analyzed from the viewpoint of kinematics. Here, we attempted to fully describe the three-dimensional (3D) kinematic motility of the growth cone. We adopted differential interference contrast (DIC) microscopic technique, which enables the high contrast live-cell imaging without phototoxic fluorescent labeling. However, the shadow-casting appearance along the shear axis had the serious disadvantages for 3D visualization and application of intensity-based image processing. To overcome these problems, we invented a novel technique that converts shadowed DIC images into self-luminous images by the Riesz transform, the multidimensional extension of the Hilbert transform. A combination of the first-order Riesz transform along the shear axis, with the second-order transform along the orthogonal axis, efficiently removed the shadow of DIC images and converted them into self-luminous images. This Riesz transform-assisted DIC (RT-DIC) imaging was applied to visualization of the 3D growth cone motility. We found that the growth cone filopodia exhibited the complex movements consisting of overall retraction, tip extension and right-screw rotation, which could be driven by the retrograde actin flow, actin polymerization and left-spiral motion of actin filaments, respectively. We further developed the automated methods for the structure and motility analysis. Filopodial structures were detected by the structure tensor calculated from RT-DIC image gradients. Motion parameters including acceleration and angular velocity were calculated by the Frenet-Serret formulas. Judging from these results, we concluded that we succeeded in complete kinematic description of the growth cone motility, suggesting that the RT-DIC imaging and its related techniques will provide a convenient but accurate method of multidimensional motility analysis.