loading page

Light Field Spatial Super-Resolution via Geometric Feature Interaction
  • +2
  • Xin Chen,
  • Yilei Chen,
  • Ping An,
  • Xinpeng Huang,
  • Chao Yang
Xin Chen
Shanghai University

Corresponding Author:wennychen@shu.edu.cn

Author Profile
Yilei Chen
Shanghai University
Author Profile
Ping An
Shanghai University
Author Profile
Xinpeng Huang
Shanghai University
Author Profile
Chao Yang
Shanghai University
Author Profile

Abstract

Light field (LF) enables high-dimensional image data representation since it can capture spatial and angular information of light rays simultaneously. The low spatial resolution caused by the limited imaging ability of the capturing equipment and the trade-off between spatial and angular resolution greatly affects the quality and application of LF images. In this letter, we propose an end-to-end LF super-resolution (SR) method via geometric feature interaction. Firstly, the low-resolution LF images are stacked in the horizontal and vertical epipolar plane image (EPI) directions and form 3D VI stacks. Then, these stacks are put into a dual-branch network, and we alternately perform 3D convolution on the viewpoint images (VIs) and EPIs by reshaping features for better feature extraction and interaction. The proposed method can fully explore the texture information and geometric consistency of the LF, and super-resolve all VIs at the same time. Experimental results on both real-world and synthetic LF datasets show that the proposed method has higher performance than other state-of-the-art methods.
19 Apr 2022Submitted to Electronics Letters
19 Apr 2022Submission Checks Completed
19 Apr 2022Assigned to Editor
28 Apr 2022Reviewer(s) Assigned
24 May 2022Review(s) Completed, Editorial Evaluation Pending
29 May 2022Editorial Decision: Revise Minor
11 Jun 20221st Revision Received
12 Jun 2022Submission Checks Completed
12 Jun 2022Assigned to Editor
12 Jun 2022Review(s) Completed, Editorial Evaluation Pending
17 Jun 2022Editorial Decision: Accept
Aug 2022Published in Electronics Letters volume 58 issue 17 on pages 654-656. 10.1049/ell2.12564