Optimization Research on Position Dynamics Skinning Animation Generation Algorithm Based on Maximal Cliques

Physics-based skinning techniques can generate soft effects such as fat jiggling, and are widely used in movies and games. Among many physics methods, Position Based Dynamics (PBD) has become the mainstream algorithm for generating physics soft skin animations due to its simplicity and speed. However, in soft skin animations, highly coupled shape matching constraints are often used to ensure the quality of soft effects. The high degree of coupling affects the decoupling quality of existing graph coloring algorithms and reduces the accuracy of soft body simulation. This paper mainly explores how to improve the decoupling quality under high coupling PBD physical models to obtain higher-precision soft body animations. To this end, the study analyzes the specific reasons for the poor decoupling effect of existing algorithms under high coupling conditions, and proposes a greedy graph coloring algorithm based on “Maximal Clique”. This algorithm uses maximal cliques to represent the highest-coupled substructures in the graph, and by prioritizing the processing of maximal cliques, it improves the decoupling quality under high coupling conditions. Experimental results show that this algorithm can effectively improve the decoupling quality and simulation accuracy of PBD physical models under high coupling conditions. Compared to existing algorithms, the accuracy of soft body skinning animation has been improved by 30% to 60%, and it effectively avoids geometric surface distortion caused by large deformations.