Feature detection and description for image matching:from hand-crafted design to deep learning

作     者：Lin Chen Franz Rottensteiner Christian Heipke Lin Chen;Franz Rottensteiner;Christian Heipke

作者机构：Institute of Photogrammetry and GeoInformation(IPI)Leibniz Universität HannoverHannoverGermany 

出 版 物：《Geo-Spatial Information Science》 (地球空间信息科学学报（英文）)

年 卷 期：2021年第24卷第1期

页      面：58-74,I0009页

核心收录：

学科分类：0303[法学-社会学] 0709[理学-地质学] 08[工学] 0708[理学-地球物理学] 080203[工学-机械设计及理论] 0705[理学-地理学] 0813[工学-建筑学] 0802[工学-机械工程] 0833[工学-城乡规划学] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：The authors would like to thank NVIDIA Corp.for donating the GPU used in this research through its GPU grant program.The first author Lin Chen would also like to thank the China Scholarship Council(CSC)for financially supporting his PhD study 

主　　题：Image matching affine shape estimation feature orientation descriptor learning image orientation 

摘      要：In feature based image matching,distinctive features in images are detected and represented by feature *** is then carried out by assessing the similarity of the descriptors of potentially conjugate *** this paper,we first shortly discuss the general ***,we review feature detection as well as the determination of affine shape and orientation of local features,before analyzing feature description in more *** the feature description review,the general framework of local feature description is presented ***,the review discusses the evolution from hand-crafted feature descriptors,***(Scale Invariant Feature Transform),to machine learning and deep learning based *** machine learning models,the training loss and the respective training data of learning-based algorithms are looked at in more detail;subsequently the various advantages and challenges of the different approaches are ***,we present and assess some current research directions before concluding the paper.