二维形状特征描述及分类识别研究进展综述

Review on Research Progress of 2D Shape Description and Classification

二维形状分类识别是计算机视觉和模式识别等领域的一个重要问题,在目标识别、图像理解等应用中发挥着重要作用。针对二维形状分类识别研究,主要从特征描述、形状分类识别、形状标准数据库三个方面综述了该方向近年来最新的研究工作。综合分析了二维形状特征表示方法,主要包括基于轮廓的、基于区域的、基于骨架的以及基于多特征融合的方法,并简要评述;介绍并分析了二维形状分类识别方法,主要包括传统机器学习分类器、集成分类器、深度学习方法等;概述了二维形状识别中常用的标准数据库;展望了二维形状识别分类研究的发展趋势。

2 D shape classification and recognition is an important issue in the fields of computer vision and pattern recognition, and plays an important role in target recognition, image understanding and other applications. In this paper, the latest research work on 2D shape classification and recognition is reviewed from three aspects:feature description, shape classification and recognition, shape standard database. First, a comprehensive analysis of 2D shape feature representation methods, including contour-based, area-based, skeleton-based and multi-feature fusion methods, is made, and a brief review is given. Then, 2D shape classification and recognition methods are introduced and analyzed, including traditional machine learning classifiers, ensemble classifiers, deep learning methods, etc. Then, the standard databases commonly used in 2D shape recognition are summarized. Finally, the development trend of 2D shape recognition classification research is prospected.