三维局部描述子综述  

3D local descriptors:a review

作　　者：方斌[1] 丁军峰 马杰[1] 明德烈[1] FANG Bin;DING Junfeng;MA Jie;MING Delie(School of Artificial Intelligence and Automation,Huazhong University of Science and Technology,Wuhan 430074,China)

机构地区：[1]华中科技大学人工智能与自动化学院,湖北武汉430074

出　　处：《华中科技大学学报（自然科学版）》2022年第11期1-15,共15页Journal of Huazhong University of Science and Technology(Natural Science Edition)

摘　　要：对三维计算机视觉领域中近三十年的局部描述子进行总结,回顾了传统三维手工局部描述符的构造方法,介绍了基于深度学习的方法.首先,针对三维手工局部特征和学习型特征,分别从局部参考坐标系和三维数据的表示方式的角度出发,对它们进行分类概述,并重点介绍部分典型方法;然后,概述了三维局部描述子的常用数据集,并统计了各数据集上现有描述子的性能;最后,探讨了三维描述子领域未来值得研究的一些问题.Local descriptors in the field of three dimension(3D) computer vision in the past three decades were summarized.Traditional methods of designing 3D manual local descriptors were reviewed,and deep learning-based methods were introduced.First,for 3D manual local features and learned features,an overview of their classification was provided from the perspectives of local reference frame system and 3D data representation,respectively,and some typical methods were highlighted.Then,common datasets of 3D local descriptors were outlined,and the performance of the existing descriptors on each dataset was statistically presented.Finally,some issues worthy of future research in the field of 3D descriptors were discussed.

关 键 词：三维 局部 手工特征 深度学习特征 参考坐标系 

分 类 号：TP37[自动化与计算机技术—计算机系统结构]