Reconstructing the local structures of Chinese ancient architecture using unsupervised depth estimation

IF 2.6 1区 艺术学 Q2 CHEMISTRY, ANALYTICAL Heritage Science Pub Date : 2024-08-30 DOI:10.1186/s40494-024-01433-9
Xiaoling Yao, Lihua Hu, Jifu Zhang
{"title":"Reconstructing the local structures of Chinese ancient architecture using unsupervised depth estimation","authors":"Xiaoling Yao, Lihua Hu, Jifu Zhang","doi":"10.1186/s40494-024-01433-9","DOIUrl":null,"url":null,"abstract":"<p>Digitalization of ancient architectures is one of the effective means for the preservation of heritage structures, with 3D reconstruction based on computer vision being a key component of such digitalization techniques. However, Chinese ancient architectures are located in mountainous areas, and existing 3D reconstruction methods fall short in restoring the local structures of these architectures. This paper proposes a self-attention-guided unsupervised single image-based depth estimation method, providing innovative technical support for the reconstruction of local structures in Chinese ancient architectures. First, an attention module is constructed based on features extracted from architectural images learned by the encoder, and then embedded into the encoder-decoder to capture the interdependencies across local features. Second, a disparity map is generated using the loss constraint network, including reconstruction matching, smoothness of the disparity, and left-right disparity consistency. Third, an unsupervised architecture based on binocular image pairs is constructed to remove any potential adverse effects due to unknown scale or estimated pose errors. Finally, with the known baseline distance and camera focal length, the disparity map is converted into the depth map to perform the end-to-end depth estimation from a single image. Experiments on the our architecture dataset validates our method, and it performs well also well on KITTI.</p>","PeriodicalId":13109,"journal":{"name":"Heritage Science","volume":"75 1","pages":""},"PeriodicalIF":2.6000,"publicationDate":"2024-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Heritage Science","FirstCategoryId":"92","ListUrlMain":"https://doi.org/10.1186/s40494-024-01433-9","RegionNum":1,"RegionCategory":"艺术学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"CHEMISTRY, ANALYTICAL","Score":null,"Total":0}
引用次数: 0

Abstract

Digitalization of ancient architectures is one of the effective means for the preservation of heritage structures, with 3D reconstruction based on computer vision being a key component of such digitalization techniques. However, Chinese ancient architectures are located in mountainous areas, and existing 3D reconstruction methods fall short in restoring the local structures of these architectures. This paper proposes a self-attention-guided unsupervised single image-based depth estimation method, providing innovative technical support for the reconstruction of local structures in Chinese ancient architectures. First, an attention module is constructed based on features extracted from architectural images learned by the encoder, and then embedded into the encoder-decoder to capture the interdependencies across local features. Second, a disparity map is generated using the loss constraint network, including reconstruction matching, smoothness of the disparity, and left-right disparity consistency. Third, an unsupervised architecture based on binocular image pairs is constructed to remove any potential adverse effects due to unknown scale or estimated pose errors. Finally, with the known baseline distance and camera focal length, the disparity map is converted into the depth map to perform the end-to-end depth estimation from a single image. Experiments on the our architecture dataset validates our method, and it performs well also well on KITTI.

Abstract Image

查看原文
分享 分享
微信好友 朋友圈 QQ好友 复制链接
本刊更多论文
利用无监督深度估计重建中国古建筑的局部结构
古建筑数字化是保护文物建筑的有效手段之一,而基于计算机视觉的三维重建技术则是这种数字化技术的重要组成部分。然而,中国的古建筑多位于山区,现有的三维重建方法无法还原这些建筑的局部结构。本文提出了一种自注意力引导的基于单幅图像的无监督深度估计方法,为中国古建筑局部结构的重建提供了创新的技术支持。首先,根据编码器从建筑图像中提取的特征构建注意力模块,然后嵌入编码器-解码器,以捕捉局部特征之间的相互依存关系。其次,利用损失约束网络生成差异图,包括重建匹配、差异平滑度和左右差异一致性。第三,构建一个基于双目图像对的无监督架构,以消除未知比例或估计姿势误差可能造成的不利影响。最后,在已知基线距离和相机焦距的情况下,将差异图转换为深度图,从而从单张图像进行端到端的深度估算。在我们的架构数据集上进行的实验验证了我们的方法,它在 KITTI 上的表现也很好。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 去求助
相关文献
Single-Phase Average Power Measurement Using Instantaneous Power Theory in a Fixed Point Processor
IF 0 2020 IEEE Calcutta Conference (CALCON)Pub Date : 2020-02-01 DOI: 10.1109/CALCON49167.2020.9106506
D. Chatterjee, S. Dalapati
Multilevel Inverter Compensation System of the Single-phase Based on Instantaneous P-instantaneous Real Power and Q-instantaneous Virtual Power Theory
IF 0 Int. J. Online Eng.Pub Date : 2015-08-31 DOI: 10.3991/ijoe.v11i7.4756
Xijuan Wang, T. Zhou, Yang Zhao, Jingxiao Feng, Jialun Zhang, Hongjun Zhang
来源期刊
Heritage Science
Heritage Science Arts and Humanities-Conservation
CiteScore
4.00
自引率
20.00%
发文量
183
审稿时长
19 weeks
期刊介绍: Heritage Science is an open access journal publishing original peer-reviewed research covering: Understanding of the manufacturing processes, provenances, and environmental contexts of material types, objects, and buildings, of cultural significance including their historical significance. Understanding and prediction of physico-chemical and biological degradation processes of cultural artefacts, including climate change, and predictive heritage studies. Development and application of analytical and imaging methods or equipments for non-invasive, non-destructive or portable analysis of artwork and objects of cultural significance to identify component materials, degradation products and deterioration markers. Development and application of invasive and destructive methods for understanding the provenance of objects of cultural significance. Development and critical assessment of treatment materials and methods for artwork and objects of cultural significance. Development and application of statistical methods and algorithms for data analysis to further understanding of culturally significant objects. Publication of reference and corpus datasets as supplementary information to the statistical and analytical studies above. Description of novel technologies that can assist in the understanding of cultural heritage.
期刊最新文献
Study on the restoration of glass slides dating back to the 1940s Greener solutions for biodeterioration of organic-media cultural heritage: where are we? Search for new materials based on chitosan for the protection of cultural heritage The diversity of bronze production technologies during the Eastern Zhou dynasty revealed by analysis of slags from the Baidian and Xincun sites in Central China Automatic defect detection in infrared thermal images of ancient polyptychs based on numerical simulation and a new efficient channel attention mechanism aided Faster R-CNN model
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
现在去查看 取消
×
提示
确定
0
微信
客服QQ
Book学术公众号 扫码关注我们
反馈
×
意见反馈
请填写您的意见或建议
请填写您的手机或邮箱
已复制链接
已复制链接
快去分享给好友吧!
我知道了
×
扫码分享
扫码分享
Book学术官方微信
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术
文献互助 智能选刊 最新文献 互助须知 联系我们:info@booksci.cn
Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。
Copyright © 2023 Book学术 All rights reserved.
ghs 京公网安备 11010802042870号 京ICP备2023020795号-1