Human Model Adaptation for Multiview Markerless Motion Capture

Abstract

An approach to automatic modeling of individual human bodies using complex shape and pose information. The aim is to address the need for human shape and pose model generation for markerless motion capture. With multi-view markerless motion capture, three-dimensional morphable models are learned from an existing database of registered body scans in different shapes and poses. We estimate the body skeleton and pose parameters from the visual hull mesh reconstructed from multiple human silhouettes. Pose variation of body shapes is implemented by the defined underlying skeleton. The shape parameters are estimated by fitting the morphable model to the silhouettes. It is done relying on extracted silhouettes only. An error function is defined to measure how well the human model fits the input data, and minimize it to get the good estimate result. Further, experiments on some data show the robustness of the method, where the body shape and the initial pose can be obtained automatically.

Virtual skeletons and digital muscles: an experimental bioarchaeological approach to the pre-Hispanic production of millstones (Tenerife, Canary Islands)

“…In order to capture static and dynamic postures, low-cost depth cameras (Kinect or RealSense, among others) have typically been used on human subjects while they perform the activity that is to be captured virtually (Tong et al, 2012;Zeng et al 2013). During the reshaping of the virtual human, it is challenging to carry out a consistent reconstruction of body size and muscles (Hasler et al, 2009;Zhang et al, 2013).…”

“…During the reshaping phase of the virtual human body, it is necessary to integrate pre-existing templates, sets of mesh, and databases of digitalized bodies in different shapes and positions. In this sense, incorporating the data generated with Kinovea may have exciting potential when building the body of virtual humans (Zhang et al, 2013;Machidon et al, 2018). In the same way, when resizing the muscles of these avatars, it would be essential to give a greater dimension to those muscles that are more used during a given physical activity.…”

Optimization and Soft Constraints for Human Shape and Pose Estimation Based on a 3D Morphable Model

“…Then, we estimated the shape parameters from multisilhouettes. For the detailed information, we refer to our former paper [19]. And the adaptation results shown in Figures 3 and 5.…”