Depth Map Super-Resolution Based on Semi-Couple Deformable Convolution Networks

Author:

Liu Botao1,Chen Kai1,Peng Sheng-Lung2ORCID,Zhao Ming1ORCID

Affiliation:

1. School of Computer Science, Yangtze University, Jingzhou 434023, China

2. Department of Creative Technologies and Product Design, National Taipei University of Business, Taipei 10051, Taiwan

Abstract

Depth images obtained from lightweight, real-time depth estimation models and consumer-oriented sensors typically have low-resolution issues. Traditional interpolation methods for depth image up-sampling result in a significant information loss, especially in edges with discontinuous depth variations (depth discontinuities). To address this issue, this paper proposes a semi-coupled deformable convolution network (SCD-Net) based on the idea of guided depth map super-resolution (GDSR). The method employs a semi-coupled feature extraction scheme to learn unique and similar features between RGB images and depth images. We utilize a Coordinate Attention (CA) to suppress redundant information in RGB features. Finally, a deformable convolutional module is employed to restore the original resolution of the depth image. The model is tested on NYUv2, Middlebury, Lu, and a Real-Sense real-world dataset created using an Intel Real-sense D455 structured-light camera. The super-resolution accuracy of SCD-Net at multiple scales is much higher than that of traditional methods and superior to recent state-of-the-art (SOTA) models, which demonstrates the effectiveness and flexibility of our model on GDSR tasks. In particular, our method further solves the problem of an RGB texture being over-transferred in GDSR tasks.

Funder

China University Industry, University and Research Innovation Fund of the Ministry of Education

Publisher

MDPI AG

Subject

General Mathematics,Engineering (miscellaneous),Computer Science (miscellaneous)

Reference56 articles.

1. Mobile augmented reality survey: From where we are to where we go;Chatzopoulos;IEEE Access,2017

2. Efficient human pose estimation from single depth images;Shotton;IEEE Trans. Pattern Anal. Mach. Intell.,2012

3. Autonomous vehicles that interact with pedestrians: A survey of theory and practice;Rasouli;IEEE Trans. Intell. Transp. Syst.,2019

4. Vision for mobile robot navigation: A survey;DeSouza;IEEE Trans. Pattern Anal. Mach. Intell.,2002

5. Guided image filtering;He;IEEE Trans. Pattern Anal. Mach. Intell.,2012

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

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