CONFERENCE PROCEEDINGS
Advanced Search
Home > Proceedings > Volume 12346 > Article
Paper
27 September 2022 Pyramid multi-view stereo with double self-attention and uncertainty awareness
Xinyu Pi, Xiaoyun Qing, Wen Li
Author Affiliations +
Proceedings Volume 12346, 2nd International Conference on Information Technology and Intelligent Control (CITIC 2022); 123460O (2022) https://doi.org/10.1117/12.2653432
Event: 2nd International Conference on Information Technology and Intelligent Control (CITIC 2022), 2022, Kunming, China
Abstract
We propose a novel model for 3D reconstruction. At present, most learning-based reconstruction methods process highresolution input images from coarse-to-fine manner, and have achieved good reconstruction results. However, these methods usually use the plane sweep volume with fixed depth hypothesis to construct the cost volume, which may miss the small boundary and the depth mutation region. We use the uncertainty estimation module to adapt to the uncertainty of depth prediction per-pixel. In addition, we introduce a double self-attention module into the feature extraction network, so that the network has the ability to capture the long-range dependent features related to depth inference tasks. Experiments were carried out on DTU datasets. The results show that our method achieves excellent performance.
© (2022) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Citation Download Citation
Xinyu Pi, Xiaoyun Qing, and Wen Li "Pyramid multi-view stereo with double self-attention and uncertainty awareness", Proc. SPIE 12346, 2nd International Conference on Information Technology and Intelligent Control (CITIC 2022), 123460O (27 September 2022); https://doi.org/10.1117/12.2653432
ACCESS THE FULL ARTICLE
ORGANIZATIONAL
Sign in with credentials provided by your organization.
INSTITUTIONAL
Select your institution to access the SPIE Digital Library.
SELECT YOUR INSTITUTION
PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.
PERSONAL SIGN IN
No SPIE Account? Create one
PURCHASE THIS CONTENT
SUBSCRIBE TO DIGITAL LIBRARY
50 downloads per 1-year subscription
Members: $195
Non-members: $335 ADD TO CART
25 downloads per 1 - year subscription
Members: $145
Non-members: $250 ADD TO CART
PURCHASE SINGLE ARTICLE
Includes PDF, HTML & Video, when available
Members: $17.00
Non-members: $21.00 ADD TO CART
PROCEEDINGS
 7 PAGES
DOWNLOAD PAPER SAVE TO MY LIBRARY
GET CITATION
Advertisement
Advertisement
RIGHTS & PERMISSIONS
Get copyright permission  Get copyright permission on Copyright Marketplace
KEYWORDS
Feature extraction
3D modeling
Image resolution
Network architectures
Computer vision technology
Image processing
Machine vision
RELATED CONTENT
Stereo matching with local cost volume refinement network
Proceedings of SPIE (January 09 2023)
Application of Random Ferns for non-planar object detection
Proceedings of SPIE (December 08 2015)
Real-time traffic sign detection based on YOLOv2
Proceedings of SPIE (October 29 2018)
Contemporary deep recurrent learning for recognition
Proceedings of SPIE (May 01 2017)
Real-time model-based vision for industrial domains
Proceedings of SPIE (August 20 1993)
Image feature matching with network flow a global optimization...
Proceedings of SPIE (January 19 2009)
Can Shape Description Be Applied To Model Matching?
Proceedings of SPIE (February 19 1988)
Subscribe to Digital Library
Receive Erratum Email Alert
Back to Top