Registration of 3D human body has been a challenging research topic for over decades. Most of the traditional human body registration methods require manual assistance, or other auxiliary information such as texture and markers. The majority of these methods are tailored for high-quality scans from expensive scanners. Following the introduction of the low-quality scans from cost-effective devices such as Kinect, the 3D data capturing of human body becomes more convenient and easier. However, due to the inevitable holes, noises and outliers in the low-quality scan, the registration of human body becomes even more challenging. To address this problem, we propose a fully automatic active registration method which deforms a high-resolution template mesh to match the low-quality human body scans. Our registration method operates on two levels of statistical shape models: (1) the first level is a holistic body shape model that defines the basic figure of human; (2) the second level includes a set of shape models for every body part, aiming at capturing more body details. Our fitting procedure follows a coarse-to-fine approach that is robust and efficient. Experiments show that our method is comparable with the state-of-the-art methods.
The modeling of accurate 3D human body is a fundamental problem for many applications such as design, animation, and virtual reality. The modeling of human body meshes is performed on a corpus of registered scans. However, the acquirement of high-quality human body meshes and registration of meshes are challenging. Current publicly available high-quality human body datasets, such as SCAPE [3], FAUST [5], TOSCA [8] are built either from costly laser scanners or need other assistance (e.g makers, texture or professional tools). With the appearance of low-cost scanners such as Kinect, it is now possible for an object, a room or even a person to be quickly scanned, modeled and tracked [12,14,27,28,30,31,44]. Nowadays human body meshes could be captured for different identities in different poses in a few minutes. However, the prevalent noises, outliers and holes in the scans acquired with low-cost scanners bring in more challenges for mesh registration.
To register the 3D scans, several 3D fitting methods are proposed [1,2,5,14,32,49]. The invertible finite volume method [14] is used to control the template tetrahedral mesh to the target point clouds. The stitched puppet model [49] adopts the DPMP algorithm which is a particle-based method to align a graphical model to target meshes. More efforts are made to perform the nonrigid ICP (iterative closest point) [1,2] which computes the Abstract Registration of 3D human body has been a challenging research topic for over decades. Most of the traditional human body registration methods require manual assistance, or other auxiliary information such as texture and markers. The majority of these methods are tailored for high-quality scans from expensive scanners. Following the introduction of the low-quality scans from cost-effective devices such as Kinect, the 3D data capturing of human body becomes more convenient and easier. However, due to the inevitable holes, noises and outliers in the low-quality scan, the registration of human body becomes even more challenging. To address this problem, we propose a fully automatic active registration method which deforms a high-resolution template mesh to match the low-quality human body scans. Our registration method operates on two levels of statistical shape models: (1) the first level is a holistic body shape model that defines the basic figure of human; (2) the second level includes a set of shape models for every body part, aiming at capturing more body details. Our fitting procedure follows a coarse-to-fine approach that is robust and efficient. Experiments show that our method is comparable with the state-of-the-art methods affine transformation at each vertex of template to allow non-rigid registration of template and scans. Although these ICP-based nonrigid registration methods demonstrate high accuracy, it is sensitive to missing data, which might lead to an erroneous fitting result. For Kinect-like scanners, due to self-occluded parts like crotch and armpit, holes and distortion on the mesh are inevitable.
To faithfully register the body scans captured from low-cost scanners, like Kinect, we present a multilevel active body registration (MABR) approach to build a watertight and high fidelity virtual human body in an automatic way. We aim to align a template mesh with the target scans acquired with Kinect as close as possible.
Here, a template mesh is the mean shape which is learned from an existing high-quality human body mesh dataset. In our method, multilevel registration is performed. In the first level, the overall template and target are roughly aligned. In the second level, a region-based registration is performed where the template is divided into 16 parts and each part is fitted to the target separately. For the main body parts where the scan is complete and full of details such as torso, legs and arms, the local affine transformation for each vertex is computed. As for impaired parts such as foot and hand, we deform the corresponding parts of the template at a coarse-grained level for completeness.
With the proposed method, we are able to automatically reconstruct high-quality 3D mesh from low-quality scans or point clouds. This technique can be employed in a variety of applications such as in virtual dressing applications to show the clothes from different stereo views and help the customers to choose the best fitting clothes. In the virtual games, the systems can generate realistic full body avatars according to rough scans of the users instantly, which benefit from algorithm’s robustness to missing data which commonly exist in scans from low-cost scanners. The approach manages to avoid the tediously manual work of building high-fidelity 3D models with professional tools and is capable of building a complete and high-quality meshes within 2 min automatically, which can be beneficial to the television production. This method may also be integrated in software as a tool for preprocessing raw scans, fil
This content is AI-processed based on open access ArXiv data.
