SCMS-Net: Self-Supervised Clustering-Based 3D Meshes Segmentation Network. (July 2023)

• Record Type: Journal Article
• Authors: Jiao, Xue; Chen, Yonggang; Yang, Xiaohui
• Is Part Of: Computer aided design. Volume 160(2023)
• Publication Date: 2023-07
• Subjects: 3D mesh segmentation -- Unsupervised segmentation -- Self-supervised learning -- Deep clustering; Computer-aided design -- Periodicals; Engineering design -- Data processing -- Periodicals; Computer graphics -- Periodicals; Conception technique -- Informatique -- Périodiques; Infographie -- Périodiques; Computer graphics; Engineering design -- Data processing; Periodicals; Electronic journals; 620.00420285
• DOI: 10.1016/j.cad.2023.103512
• Languages: English
• ISSNs: 0010-4485

Abstract

Abstract: The superior performance of deep learning in different domains has sparked significant interest in its applicability to 3D computer graphics. Deep learning has become the dominant technical architecture in current 3D mesh segmentation. However, learning-based 3D segmentation methods usually rely on high-quality training datasets, which are not readily available in practical applications. How to segment 3D meshes without exhaustive label annotations remains a challenging problem, especially in the context of deep learning. As a subset of unsupervised learning methods, self-supervised learning offers a promising learning paradigm for unlabeled 3D mesh segmentation. In this paper, we introduce a self-supervised clustering-Based network specifically for the segmentation of label-free 3D meshes. Our self-supervised clustering-based 3D mesh segmentation network (SCMS-Net) employs a two-branch architecture to learn effective feature representation. The two branches are unified into an end-to-end framework using a self-supervised strategy. Finally, the label predictions of the parts are generated by iterative clustering. We conducted ablation studies and comparative experiments on a standard benchmark to demonstrate the effectiveness of our approach. Highlights: This is the first work to explore the self-supervised clustering methods for mesh segmentation. We use a two-branch network to integrate geometric and structural information. The segmentation loss function consists of the clustering and reconstruction loss.