Helical computed tomography (HCT) has become the preferred protocol in routine clinical applications. The advantages of HCT include its capability of scanning a complete anatomical volume in a single breath-hold, and the capability of generating images at arbitrary locations. However, studies have indicated various HCT-related image artifacts in 3D applications. In this paper, we perform a detailed analysis on the cause of one of the artifacts in the 3D shaded surface display (SSD): the "grooving" artifact. We developed an analytical model to show that the artifact is induced by the angular-dependent edge response of the HCT reconstruction. The accuracy of the model is demonstrated by both computer simulations and phantom experiments. Based on the model, we present several correction algorithms to combat the "grooving" artifacts. Computer simulations and phantom experiments have shown the effectiveness and thoroughness of the correction