Research On Cloth Simulation Modeling Integrated With Machine Learning Algorithm

Cloth simulation has a long research history and important research significance in the fields of physical simulation,computer games,virtual reality and so on.The fabrics in real life have characteristics such as abundant folds and different shapes.How to efficiently use the computer to simulate these fabric characteristics has become a difficult point in fabric simulation research.In cloth simulation modeling,the accuracy of cloth simulation and the speed of cloth simulation often restrict each other.At present,there are some traditional cloth simulation methods that generally can only take into account one of the two,it is difficult to achieve a balance,so how to find a The method of taking into account both simulation accuracy and simulation speed to achieve a certain degree of balance is the focus of research in cloth simulation technology.The spring-mass model is a grid model commonly used in cloth simulation modeling.In the process of using the spring-mass model for cloth simulation,it is necessary to calculate and predict the position of the mass at the next moment.Most of them use physical-based methods for particle position prediction,which have some shortcomings such as complicated calculation,long time consumption and poor real-time performance.In view of the above problems,this paper starts from the method of calculating the position of the mass point,improves the traditional physical-based method,and studies a faster and more accurate cloth modeling method,which provides a numerical calculation method for cloth simulation that can take into account the calculation accuracy and calculation efficiency.Specific work includes the following:?1?Aiming at the shortcomings of complex calculation based on physical method,timeconsuming and poor real-time performance,a cloth layered modeling method combining random forest model is proposed.A physical-based method is used to calculate the initial position of each mass point,and each mass point is connected to form the most initial horizontal cloth.Then use the regression algorithm of the random forest model to infer the position of the particles at the next level of cloth,use the 3^1/2 subdivision method to connect the particles,and then generate a stable cloth grid through the edge flip operation,and repeat the above process until a satisfactory animation effect is produced.?2?The random forest model feels like a black box and cannot control the internal operation of the model.It can only be tried between different parameters and random seeds,and the maneuverability is poor.Compared with the random forest algorithm,the neural network algorithm is more maneuverable,and when an appropriate network model is constructed,the computational performance of the neural network algorithm is also superior to other traditional methods.Therefore,the neural network model is used to predict the position of the particles,and the cloth is simulated at a more detailed level.