A dynamic clothing simulation method based on human body mixed bounding box

Abstract

The invention relates to the technical field of dynamic human clothing simulation, in particular to a dynamic clothing simulation method based on a human body mixed bounding box. This method first uses the MCASG graph algorithm and the K-means clustering algorithm to perform primary segmentation and secondary segmentation on the complete human body model to obtain the head, upper arm, lower arm, upper leg, lower leg and torso of the human body model. . Second, bounding boxes that fit the shape of different parts of the human body are selected, such as spherical bounding boxes for buttocks and female breasts, cylindrical bounding boxes for arms and legs, and elliptical cylindrical bounding boxes for torso. Finally, the clothing is modeled by using the spring particle model based on the triangular surface, and the collision detection and response are carried out to realize the realistic real-time clothing simulation. The dynamic clothing simulation method based on the mixed bounding box of the human body described in the present invention can effectively improve the collision detection speed of the clothing and the human body model, and shorten the collision processing time.

Description

technical field [0001] The invention relates to the technical field of dynamic human clothing simulation, in particular to a dynamic clothing simulation method based on a human body mixed bounding box. Background technique [0002] The simulation of human body clothing based on physical model has been a hot research content and challenging topic in clothing CAD and computer graphics. It requires cross-application of knowledge in the fields of computer, mathematics, physics, clothing engineering, etc., focusing on real, real-time, and lightweight generation of the actual effects of fabric and clothing movement and deformation, in virtual fittings, human body simulation, movie games, etc. It has been widely used in the field. [0003] In the process of dynamic human clothing simulation, it mainly involves the effective physical model construction of clothing, collision detection between human body and clothing, etc. For the physical simulation of clothing, the first problem ...

AI Technical Summary

Problems solved by technology

However, the existing bounding box technology often has defects such as too many bounding boxes, large shape errors, complex calculations, and low collision detection efficiency.

For example: Weller et al. used the inscribed ball tree to detect the collision and cut-in of rigid bodies, but the number of spherical bounding boxes at the bottom of the spherical tree is often too large; Bradshaw et al. used methods such as fusion, explosion, and expansion to reduce the number of spheres. It is isotropic. When expressing sharp or flat features, it is prone to large out-of-shape errors. Choosing an ellipsoid as a bounding volume can effectively solve this problem; Bischoff et al. use seed points to construct the initial ellipsoid method, Decomposing the model with an ellipsoid, but simply expressing it with an ellipsoid, there is a problem of computational complexity, which cannot effectively simplify the computational time complexity

Images