Point cloud segmentation is an essential task in three-dimensional (3D) vision and intelligence. It is a critical step in understanding 3D scenes with a variety of applications. With the rapid development of 3D scanning devices, point cloud data have become increasingly available to researchers. Recent advances in deep learning are driving advances in point cloud segmentation research and applications. This paper presents a comprehensive review of recent progress in point cloud segmentation for understanding 3D indoor scenes. First, we present public point cloud datasets, which are the foundation for research in this area. Second, we briefly review previous segmentation methods based on geometry. Then, learning-based segmentation methods with multi-views and voxels are presented. Next, we provide an overview of learning-based point cloud segmentation, ranging from semantic segmentation to instance segmentation. Based on the annotation level, these methods are categorized into fully supervised and weakly supervised methods. Finally, we discuss open challenges and research directions in the future.
With the increase in dental patient numbers and the ongoing digital transformation of dental hospitals, tooth segmentation has become increasingly crucial for the digital diagnosis, design, treatment, and customized appliance manufacturing of orthodontics, oral implant surgery, and prosthodontics. This study aims to adapt the Segment Anything Model (SAM) to the Tooth segmentation task for precise tooth segmentation performance. In this study, a novel tooth segmentation method-Tooth-ASAM-that harnesses the power of SAM was introduced. An adapter-based image encoder and mask decoder specifically tailored for adapting SAM to tooth images were designed. The proposed method was evaluated through rigorous evaluation of multimodal tooth images-including Cone Beam Computed Tomography (CBCT) images, panoramic X-rays, and natural teeth images captured by a micro-camera. The experimental results unequivocally show that Tooth-ASAM achieved remarkable performances across all four datasets, excelling in key metrics like the Dice coefficient, IoU, HD95, and ASSD. Furthermore, the proposed Tooth-ASAM delivered perceptually more accurate segmentation results than the state-of-the-art methods on the four tooth datasets. This research demonstrates that precise tooth segmentation performances were obtained by applying SAM and adaptation training strategy, making it highly suitable for clinical applications in orthodontics, oral implant surgery, and prosthodontics.
16 April 2024 in Multimedia Tools and Applications