A dual-branch stereo image super-resolution reconstruction network integrating internal and external view information
Measurement Science and Technology · 29.05.2026 · 37 (22) · 225403
DOI: 10.1088/1361-6501/ae70c8 ↗Kurz gefasst
Fachartikel in der Zeitschrift Measurement Science and Technology, online erschienen am 02.06.2026. Verfasst von Zhenghui Xu und 2 weiteren Autor:innen. Die Arbeit nennt 39 Literaturangaben.
Abstract (Originalsprache)
The stereo image super-resolution reconstruction (SR) technology can effectively leverage the complementary information across multiple-view images, enabling precise reconstruction of image details. However, in stereo images, not only do rich-texture regions exist, but vast low-texture regions are also prevalent, which significantly increases the challenge of accurately extracting cross-view complementary information. To overcome this challenge, a dual-branch stereo image SR network integrating internal and external view information is proposed. First, a dual-branch feature extraction module (DFEM) is proposed. By leveraging convolution kernels with different receptive fields to simultaneously extract the high-frequency and low-frequency information from the input image, thereby enhancing the feature representation. Then, an internal and external view information fusion module (IEVFM) is proposed to capture features from individual views and across multiple views, achieving more accurate extraction of complementary information. Next, a feature refinement module is designed to filter and refine the fused features, extracting more useful high-frequency information. Finally, by applying upsampling operations to the refined features, the SR of low-resolution (LR) stereo images is achieved. To validate the effectiveness of the proposed method, experiments are performed on Flickr1024, KITTI2012, KITTI2015, and Middlebury2014 datasets. The experimental results demonstrate that the proposed method can more efficiently mine and utilize the complementary information in LR stereo images, thereby improving the quality of their SR.
Zitieren
APA
Xu, Z., Wu, J., & Wang, Y. (2026). A dual-branch stereo image super-resolution reconstruction network integrating internal and external view information. Measurement Science and Technology, 37(22), 225403. https://doi.org/10.1088/1361-6501/ae70c8
Vancouver
Xu Z, Wu J, Wang Y. A dual-branch stereo image super-resolution reconstruction network integrating internal and external view information. Measurement Science and Technology. 2026;37(22):225403. doi:10.1088/1361-6501/ae70c8
BibTeX
@article{Xu2026,
author = {Xu, Zhenghui and Wu, Jiayi and Wang, Yunming},
title = {A dual-branch stereo image super-resolution reconstruction network integrating internal and external view information},
journal = {Measurement Science and Technology},
year = {2026},
volume = {37},
number = {22},
pages = {225403},
doi = {10.1088/1361-6501/ae70c8}
}
EUVIMED