Disentangling Light Fields for Super-Resolution and Disparity Estimation

被引:141
作者
Wang, Yingqian [1 ]
Wang, Longguang [1 ]
Wu, Gaochang [2 ]
Yang, Jungang [1 ]
An, Wei [1 ]
Yu, Jingyi [3 ]
Guo, Yulan [1 ]
机构
[1] Natl Univ Def Technol, Coll Elect Sci & Technol, Changsha 410073, Hunan, Peoples R China
[2] Northeastern Univ, State Key Lab Synthet Automat Proc Ind, Shenyang 110819, Peoples R China
[3] ShanghaiTech Univ, Sch Informat Sci & Technol, Pudong 201210, Peoples R China
关键词
Light field image processing; feature disentangling; image super-resolution; view synthesis; disparity estimation; EPIPOLAR GEOMETRY; NETWORK; DEPTH; SHAPE;
D O I
10.1109/TPAMI.2022.3152488
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Light field (LF) cameras record both intensity and directions of light rays, and encode 3D scenes into 4D LF images. Recently, many convolutional neural networks (CNNs) have been proposed for various LF image processing tasks. However, it is challenging for CNNs to effectively process LF images since the spatial and angular information are highly inter-twined with varying disparities. In this paper, we propose a generic mechanism to disentangle these coupled information for LF image processing. Specifically, we first design a class of domain-specific convolutions to disentangle LFs from different dimensions, and then leverage these disentangled features by designing task-specific modules. Our disentangling mechanism can well incorporate the LF structure prior and effectively handle 4D LF data. Based on the proposed mechanism, we develop three networks (i.e., DistgSSR, DistgASR and DistgDisp) for spatial super-resolution, angular super-resolution and disparity estimation. Experimental results show that our networks achieve state-of-the-art performance on all these three tasks, which demonstrates the effectiveness, efficiency, and generality of our disentangling mechanism. Project page: https://yingqianwang.github.io/DistgLF/.
引用
收藏
页码:425 / 443
页数:19
相关论文
共 82 条
  • [1] Alain M, 2018, IEEE IMAGE PROC, P2501, DOI 10.1109/ICIP.2018.8451162
  • [2] Alain M, 2017, IEEE INT WORKSH MULT
  • [3] Light field intrinsics with a deep encoder-decoder network
    Alperovich, Anna
    Johannsen, Ole
    Strecke, Michael
    Goldluecke, Bastian
    [J]. 2018 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2018, : 9145 - 9154
  • [4] [Anonymous], 2013, Vision, Modelling and Visualization (VMV)
  • [5] The Light Field Camera: Extended Depth of Field, Aliasing, and Superresolution
    Bishop, Tom E.
    Favaro, Paolo
    [J]. IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2012, 34 (05) : 972 - 986
  • [6] Geometric Calibration of Micro-Lens-Based Light Field Cameras Using Line Features
    Bok, Yunsu
    Jeon, Hae-Gon
    Kweon, In So
    [J]. IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2017, 39 (02) : 287 - 300
  • [7] Chen JX, 2021, AAAI CONF ARTIF INTE, V35, P1009
  • [8] Light Field Super-Resolution with Zero-Shot Learning
    Cheng, Zhen
    Xiong, Zhiwei
    Chen, Chang
    Liu, Dong
    Zha, Zheng-Jun
    [J]. 2021 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, CVPR 2021, 2021, : 10005 - 10014
  • [9] Learning a Deep Convolutional Network for Image Super-Resolution
    Dong, Chao
    Loy, Chen Change
    He, Kaiming
    Tang, Xiaoou
    [J]. COMPUTER VISION - ECCV 2014, PT IV, 2014, 8692 : 184 - 199
  • [10] Egiazarian K, 2015, EUR SIGNAL PR CONF, P2849, DOI 10.1109/EUSIPCO.2015.7362905