Simulating 3D fabric drape models with diversity based on point cloud random resampling and blender

The drape variability, as one of the inherent characteristics of fabrics, is characterized by the fact that multiple repetitions of the same fabric sample under the same experimental conditions will yield different results. Nevertheless, existing fabric drape simulation studies do not yield diversified simulation results. It has been pointed out that this feature makes the drape models obtained from fabric simulation never accurately mimic the real fabric. To address this problem, random point cloud resampling and Blender were used in this study to simulate various drape models under the same simulating condition. Firstly, a dense 2D point cloud with a circular shape was generated. Subsequently, a random sampling method was utilized to resample the dense point clouds into sparse ones with a fixed number of vertices. Then, the resampled sparse point clouds were encapsulated into a triangle mesh with a circular shape. Finally, the mesh was imported into Blender to simulate the natural drape of a circular fabric under gravity over a circular supporting disk. The drape indicators of the 3D fabric drape simulation results were extracted and analyzed. The results show that the method proposed in this study can be used to obtain a rich variety of 3D drape models in multiple repetitions of simulations under the same simulation parameters. There is a high similarity between the simulated and real fabric drape models in terms of variation coefficients of the drape indicators. In order to apply the simulation method proposed in this study, six sets of 3D drape models were simulated, so as to render and form a fabric drape image classification dataset. A deep learning classification model was trained with this dataset to identify fabric drape images. The results show that the fabric drape classification accuracy of the model could reach 98%. The classification model could be used to extract the feature of fabric drape, and faciliate the search of fabric drape models or fabrics.