CPC G06T 17/20 (2013.01) [G06T 7/55 (2017.01); G06T 7/80 (2017.01); G06T 2207/30196 (2013.01)] | 9 Claims |
1. A three-dimensional skeleton generation method using calibration based on a joint acquired from a plurality of cameras, the three-dimensional skeleton generation method comprising:
(a) acquiring a multiview color-depth video of a plurality of viewpoints;
(b) generating a three-dimensional skeleton of each viewpoint from a color-depth video of each viewpoint, and generating a joint of the skeleton of each viewpoint as a feature point;
(c) performing extrinsic calibration for optimizing an extrinsic parameter by using the joint of the skeleton of each viewpoint; and
(d) aligning and integrating the three-dimensional skeleton of each viewpoint by using the extrinsic parameter,
wherein the step (c) includes optimizing a transformation parameter such that an error between an actual coordinate (Xref) of a point cloud of a reference coordinate system and a transformation coordinate (Xi′) by the transformation parameter is minimized, in which the optimization is repeatedly performed by updating a next coordinate transformation parameter Pn+from a current coordinate transformation parameter Pn by the following equation:
where α is a preset constant, and ∂fError/∂Pn is a partial differentiation of fError with respect to the transformation parameter, and fError is an error function of the coordinate (Xref) and the transformation coordinate (Xi′) by the transformation parameter.
|