Deep 3D reconstruction:methods,data,and challenges

作     者：Caixia LIU Dehui KONG Shaofan WANG Zhiyong WANG Jinghua LI Baocai YIN Caixia LIU;Dehui KONG;Shaofan WANG;Zhiyong WANG;Jinghua LI;Baocai YIN

作者机构：Beijing Key Laboratory of Multimedia and Intelligent Software TechnologyBeijing Institute of Artificial IntelligenceFaculty of Information TechnologyBeijing University of TechnologyBeijing 100124China Multimedia LaboratorySchool of Computer ScienceUniversity of SydneySydney NSW 2006Australia 

出 版 物：《Frontiers of Information Technology & Electronic Engineering》 (信息与电子工程前沿（英文版）)

年 卷 期：2021年第22卷第5期

页      面：652-672页

核心收录：

学科分类：081203[工学-计算机应用技术] 08[工学] 0835[工学-软件工程] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：Project supported by the National Natural Science Foundation of China(Nos.61772049,61632006,61876012,U19B2039,and 61906011) the Beijing Natural Science Foundation of China(No.4202003) 

主　　题：Deep learning models Three-dimensional reconstruction Recurrent neural network Deep autoencoder Generative adversarial network Convolutional neural network 

摘      要：Three-dimensional(3D)reconstruction of shapes is an important research topic in the fields of computer vision,computer graphics,pattern recognition,and virtual *** 3D reconstruction methods usually suffer from two bottlenecks:(1)they involve multiple manually designed states which can lead to cumulative errors,but can hardly learn semantic features of 3D shapes automatically;(2)they depend heavily on the content and quality of images,as well as precisely calibrated *** a result,it is difficult to improve the reconstruction accuracy of those methods.3D reconstruction methods based on deep learning overcome both of these bottlenecks by automatically learning semantic features of 3D shapes from low-quality images using deep ***,while these methods have various architectures,in-depth analysis and comparisons of them are unavailable so *** present a comprehensive survey of 3D reconstruction methods based on deep ***,based on different deep learning model architectures,we divide 3D reconstruction methods based on deep learning into four types,recurrent neural network,deep autoencoder,generative adversarial network,and convolutional neural network based methods,and analyze the corresponding methodologies ***,we investigate four representative databases that are commonly used by the above methods in ***,we give a comprehensive comparison of 3D reconstruction methods based on deep learning,which consists of the results of different methods with respect to the same database,the results of each method with respect to different databases,and the robustness of each method with respect to the number of ***,we discuss future development of 3D reconstruction methods based on deep learning.