Light Field Angular Super-Resolution via Spatial-Angular Correlation Extracted by Deformable Convolutional Network

Light Field Angular Super-Resolution (LFASR) addresses the issue where Light Field (LF) images can not simultaneously achieve both high spatial and angular resolution due to the limited resolution of optical sensors. Since Spatial-Angular Correlation (SAC) features are closely related to the structu...

Full description

Saved in:
Bibliographic Details
Main Authors: Daichuan Li, Rui Zhong, Yungang Yang
Format: Article
Language:English
Published: MDPI AG 2025-02-01
Series:Sensors
Subjects:
Online Access:https://www.mdpi.com/1424-8220/25/4/991
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849722261843476480
author Daichuan Li
Rui Zhong
Yungang Yang
author_facet Daichuan Li
Rui Zhong
Yungang Yang
author_sort Daichuan Li
collection DOAJ
description Light Field Angular Super-Resolution (LFASR) addresses the issue where Light Field (LF) images can not simultaneously achieve both high spatial and angular resolution due to the limited resolution of optical sensors. Since Spatial-Angular Correlation (SAC) features are closely related to the structure of LF images, its accurate and complete extraction is crucial for the quality of LF images reconstructed by the LFASR method based on Deep Neural Networks (DNNs). In low-angular resolution LF images, SAC features must be extracted from a limited number of pixels that are at a great distance from each other and exhibit strong correlations. However, existing LFASR methods based on DNNs fail to extract SAC features accurately and completely. Due to the limited receptive field, methods based on regular Convolutional Neural Networks (CNNs) are unable to capture SAC features from distant pixels, leading to incomplete SAC feature extraction. On the other hand, methods based on large convolution kernels and attention mechanisms use an excessive number of pixels to extract SAC features, resulting in insufficient accuracy in extracted SAC features. To solve this problem, we introduce Deformable Convolutional Network (DCN), which adaptively changes the position of limited sampling point using offsets, so as to extract SAC from distant pixels. In addition, in order to make the offset of DCN more accurate and further improve the accuracy of SAC features, we also propose a Multi-Maximum-Offsets Fusion DCN (MMOF-DCN). MMOF-DCN can reduce the exploration range of finding the desired offset, thereby improving the offset finding efficiency. Experiment results show that our proposed method has advantages in real-world dataset and synthetic dataset. The PSNR value in synthetic dataset which have large disparity is improved by 0.45 dB compared to existing methods.
format Article
id doaj-art-e1e9e859597b43529e87a1ad6838cf78
institution DOAJ
issn 1424-8220
language English
publishDate 2025-02-01
publisher MDPI AG
record_format Article
series Sensors
spelling doaj-art-e1e9e859597b43529e87a1ad6838cf782025-08-20T03:11:22ZengMDPI AGSensors1424-82202025-02-0125499110.3390/s25040991Light Field Angular Super-Resolution via Spatial-Angular Correlation Extracted by Deformable Convolutional NetworkDaichuan Li0Rui Zhong1Yungang Yang2School of Computer Science, Central China Normal University, Wuhan 430079, ChinaSchool of Computer Science, Central China Normal University, Wuhan 430079, ChinaSchool of Computer Science, Central China Normal University, Wuhan 430079, ChinaLight Field Angular Super-Resolution (LFASR) addresses the issue where Light Field (LF) images can not simultaneously achieve both high spatial and angular resolution due to the limited resolution of optical sensors. Since Spatial-Angular Correlation (SAC) features are closely related to the structure of LF images, its accurate and complete extraction is crucial for the quality of LF images reconstructed by the LFASR method based on Deep Neural Networks (DNNs). In low-angular resolution LF images, SAC features must be extracted from a limited number of pixels that are at a great distance from each other and exhibit strong correlations. However, existing LFASR methods based on DNNs fail to extract SAC features accurately and completely. Due to the limited receptive field, methods based on regular Convolutional Neural Networks (CNNs) are unable to capture SAC features from distant pixels, leading to incomplete SAC feature extraction. On the other hand, methods based on large convolution kernels and attention mechanisms use an excessive number of pixels to extract SAC features, resulting in insufficient accuracy in extracted SAC features. To solve this problem, we introduce Deformable Convolutional Network (DCN), which adaptively changes the position of limited sampling point using offsets, so as to extract SAC from distant pixels. In addition, in order to make the offset of DCN more accurate and further improve the accuracy of SAC features, we also propose a Multi-Maximum-Offsets Fusion DCN (MMOF-DCN). MMOF-DCN can reduce the exploration range of finding the desired offset, thereby improving the offset finding efficiency. Experiment results show that our proposed method has advantages in real-world dataset and synthetic dataset. The PSNR value in synthetic dataset which have large disparity is improved by 0.45 dB compared to existing methods.https://www.mdpi.com/1424-8220/25/4/991light fieldangular super-resolutionoptical sensorsreconstructdeep neural network
spellingShingle Daichuan Li
Rui Zhong
Yungang Yang
Light Field Angular Super-Resolution via Spatial-Angular Correlation Extracted by Deformable Convolutional Network
Sensors
light field
angular super-resolution
optical sensors
reconstruct
deep neural network
title Light Field Angular Super-Resolution via Spatial-Angular Correlation Extracted by Deformable Convolutional Network
title_full Light Field Angular Super-Resolution via Spatial-Angular Correlation Extracted by Deformable Convolutional Network
title_fullStr Light Field Angular Super-Resolution via Spatial-Angular Correlation Extracted by Deformable Convolutional Network
title_full_unstemmed Light Field Angular Super-Resolution via Spatial-Angular Correlation Extracted by Deformable Convolutional Network
title_short Light Field Angular Super-Resolution via Spatial-Angular Correlation Extracted by Deformable Convolutional Network
title_sort light field angular super resolution via spatial angular correlation extracted by deformable convolutional network
topic light field
angular super-resolution
optical sensors
reconstruct
deep neural network
url https://www.mdpi.com/1424-8220/25/4/991
work_keys_str_mv AT daichuanli lightfieldangularsuperresolutionviaspatialangularcorrelationextractedbydeformableconvolutionalnetwork
AT ruizhong lightfieldangularsuperresolutionviaspatialangularcorrelationextractedbydeformableconvolutionalnetwork
AT yungangyang lightfieldangularsuperresolutionviaspatialangularcorrelationextractedbydeformableconvolutionalnetwork