Three-Dimensional-Slice-Super-Resolution-Net: A Fast Few Shooting Learning Model for 3D Super-Resolution Using Slice-Up and Slice-Reconstruction

Author:

Lin Hongbin12,Xu Qingfeng134,Xu Handing134,Xu Yanjie134,Zheng Yiming15,Zhong Yubin2,Nie Zhenguo134

Affiliation:

1. Tsinghua University Department of Mechanical Engineering, , Beijing 100084 , China ;

2. Guangzhou University School of Mathematics and Information Science, , Guangzhou 510006 , China

3. Tsinghua University State Key Laboratory of Tribology in Advanced Equipment, , Beijing 100084 , China ;

4. Tsinghua University Beijing Key Laboratory of Precision/Ultra-Precision Manufacturing Equipments and Control, , Beijing 100084 , China

5. Beijing Union University College of Urban Rail Transit and Logistics, , Beijing 100101 , China

Abstract

Abstract 3D modeling accurately depicts the physical world but typically requires substantial data acquisition resources and significant storage space. We introduce a novel three-dimensional slice-reconstruction model (3DSR) to address these challenges. This 3D data super-resolution model leverages low-resolution 3D data as input to generate high-resolution results promptly and accurately, reducing the time and storage required to create detailed 3D models. To enhance the computational efficiency and accuracy of deep learning models, the 3D data are partitioned into multiple slices. The 3DSR processes each slice into a high-resolution 2D image, which is then reassembled into high-resolution 3D data. Our slice-up method and slice-reconstruction technique are specifically designed to preserve the primary characteristics of the 3D data. We employ a pre-trained deep 2D convolutional neural network to expand the resolution of the 2D image, resulting in excellent performance. This approach reduces the time required for training deep learning models and enhances computational efficiency during the resolution improvement process. Importantly, our model can deliver superior performance even when trained on fewer data.

Funder

National Natural Science Foundation of China

Publisher

ASME International

Subject

Industrial and Manufacturing Engineering,Computer Graphics and Computer-Aided Design,Computer Science Applications,Software

Reference30 articles.

1. Shape Inpainting Using 3D Generative Adversarial Network and Recurrent Convolutional Networks;Wang,2017

2. Dense 3D Object Reconstruction From a Single Depth View;Yang;IEEE Trans. Pattern. Anal. Mach. Intell.,2018

3. Detailed 3D Human Body Reconstruction From Multi-View Images Combining Voxel Super-Resolution and Learned Implicit Representation;Li;Appl. Intell.,2022

4. Yolov3: An Incremental Improvement;Redmon,2018

5. U-Net: Convolutional Networks for Biomedical Image Segmentation;Ronneberger,2015

Cited by 1 articles. 订阅此论文施引文献 订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献

1. Multi-Modal Machine Learning in Engineering Design: A Review and Future Directions;Journal of Computing and Information Science in Engineering;2023-11-24

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3