Global Non-rigid Registration And Reconstruction Of Deformable Objects

Dynamic 3D reconstruction is an active research topic in computer graphics and computer vision,which tries to recover dynamic scenes by capturing videos using multiple cameras or a single camera.With the appearance of commodity depth cameras,e.g.,Microsoft Kinect,it is easier and cheaper to reconstruct the shape and texture of a 3D scene using a single RGB-D camera.This has wide applicability in 3D printing,gaming,and movie production.However,when deformable objects are concerned,reconstruction results obtained using KinectFusion has serious drifting artifacts.Moreover,the captured depth point clouds usually contain a lot of noise and outliers.Hence,it remains a huge challenge to reconstruct dynamic 3D scenes using a single RGB-D camera.We present a novel global registration method for deformable objects captured using a single RGB-D camera.Our algorithm allows objects to undergo large non-rigid deformations,and achieves high quality results without constraining the actor's pose or camera motion.We compute the deformations of all the scans simultaneously by optimizing a global alignment problem to avoid the well-known loop closure problem,and use an as-rigid-as-possible constraint to eliminate the shrinkage problem of the deformed model.To attack large scale problems,we design a coarse-to-fine multiresolution scheme,which also avoids the optimization being trapped into local minima.The proposed method is evaluated on public datasets and real datasets captured by an RGB-D sensor.Experimental results demonstrate that the proposed method obtains better results on geometry alignment than the state-of-the-art methods,and is more robust to noisy and outliers.