3D Skeleton Extraction from Volume Data Based on Normalized Gradient Vector Flow

Abstract

Skeleton extraction and visualization of 3D reconstructed target objects from multiple views continues to be a major challenge in terms of providing intuitive and uncluttered images that allow the users to understand their data. This paper presents a three-dimensional skeleton extraction technique of deformable objects based on a normalized gradient vector flow in order to analyze and visualize its characteristics. 3D deformable objects are reconstructed by an image based visual hull technique from known extrinsic and intrinsic camera parameters and silhouettes which are extracted from each camera. Our 3D skeleton extraction methodology employs the normalized gradient vector flow which is a vector diffusion approach based on partial differential equations. The euclidean distance of the magnitude of a normalized gradient vector flow is used to extract the medial axis of volume data. A markerless 3D skeletonization of reconstructed objects from multiple images might be applied to retrieve the 3D model or correct the 3D motion of the target objects.