Mesh Segmentation with Enhanced Shape Perception for CAD Models

The primitive segmentation and fitting of CAD meshes are important issues in the field of computer-aided design. The segmentation results of existing methods have large errors with the surface primitives. In this regard, a CAD mesh segmentation method with enhanced shape perception is proposed. The segmentation region is generated by expanding from seed surfaces, while surface primitives are fitted synchronously, thereby improving the consistency between the segmentation region and the surface primitives. By using the distance from the face to the primitive as the criterion, combined with an adaptive threshold controlling method, noisy models can be robustly handled. For blending surfaces between primitives, given that their curve skeleton can better express the topology between adjacent primitives, a new segmentation method is proposed. Comparative experiments on segmentation and fitting are carried out on 146 models from the ABC dataset. The results show that the proposed method improves the mIoU by 18% and reduces the RMSE by 87% on average.