The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. ex. Some numerals are expressed as "XNUMX".
Copyrights notice
The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. Copyrights notice
pandangan teks lengkap
112
Medan cahaya, yang bersamaan dengan set padat imej berbilang paparan, mempunyai pelbagai aplikasi seperti anggaran kedalaman dan paparan 3D. Salah satu masalah penting dalam aplikasi medan cahaya ialah interpolasi medan cahaya, iaitu, interpolasi pandangan. Ketepatan interpolasi dipertingkatkan dengan mengeksploitasi sifat sedia ada medan cahaya. Satu contoh ialah imej satah epipolar (EPI), yang merupakan subset 2D medan cahaya 4D, terdiri daripada banyak garisan dan garisan ini mempunyai cerun yang hampir sama di kawasan setempat. Struktur ini mendorong perwakilan yang jarang dalam domain frekuensi, di mana kebanyakan tenaga berada pada garis yang melalui asal. Atas dasar pemerhatian ini, kami mencadangkan kumpulan jarang yang sesuai untuk medan cahaya untuk mengeksploitasi struktur garisannya sepenuhnya untuk interpolasi. Khususnya, kami mereka bentuk kumpulan arah dalam domain transformasi Fourier diskret (DFT) supaya kumpulan boleh mewakili kepekatan tenaga, dan dengan itu kami merumuskan masalah interpolasi LF sebagai laso kumpulan bertindih. Kami juga memperkenalkan beberapa teknik untuk meningkatkan ketepatan interpolasi seperti menggunakan fungsi tetingkap, menentukan berat kumpulan, mengembangkan blok pemprosesan dan menggabungkan blok. Keputusan eksperimen kami menunjukkan bahawa kaedah yang dicadangkan boleh mencapai kualiti yang lebih baik atau setanding berbanding dengan kaedah interpolasi LF yang canggih seperti kaedah berasaskan rangkaian neural konvolusi (CNN).
Shu FUJITA
Nagoya University
Keita TAKAHASHI
Nagoya University
Toshiaki FUJII
Nagoya University
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Salinan
Shu FUJITA, Keita TAKAHASHI, Toshiaki FUJII, "Good Group Sparsity Prior for Light Field Interpolation" in IEICE TRANSACTIONS on Fundamentals,
vol. E103-A, no. 1, pp. 346-355, January 2020, doi: 10.1587/transfun.2018EAP1175.
Abstract: A light field, which is equivalent to a dense set of multi-view images, has various applications such as depth estimation and 3D display. One of the essential problems in light field applications is light field interpolation, i.e., view interpolation. The interpolation accuracy is enhanced by exploiting an inherent property of a light field. One example is that an epipolar plane image (EPI), which is a 2D subset of the 4D light field, consists of many lines, and these lines have almost the same slope in a local region. This structure induces a sparse representation in the frequency domain, where most of the energy resides on a line passing through the origin. On the basis of this observation, we propose a group sparsity prior suitable for light fields to exploit their line structure fully for interpolation. Specifically, we designed the directional groups in the discrete Fourier transform (DFT) domain so that the groups can represent the concentration of the energy, and we thereby formulated an LF interpolation problem as an overlapping group lasso. We also introduce several techniques to improve the interpolation accuracy such as applying a window function, determining group weights, expanding processing blocks, and merging blocks. Our experimental results show that the proposed method can achieve better or comparable quality as compared to state-of-the-art LF interpolation methods such as convolutional neural network (CNN)-based methods.
URL: https://global.ieice.org/en_transactions/fundamentals/10.1587/transfun.2018EAP1175/_p
Salinan
@ARTICLE{e103-a_1_346,
author={Shu FUJITA, Keita TAKAHASHI, Toshiaki FUJII, },
journal={IEICE TRANSACTIONS on Fundamentals},
title={Good Group Sparsity Prior for Light Field Interpolation},
year={2020},
volume={E103-A},
number={1},
pages={346-355},
abstract={A light field, which is equivalent to a dense set of multi-view images, has various applications such as depth estimation and 3D display. One of the essential problems in light field applications is light field interpolation, i.e., view interpolation. The interpolation accuracy is enhanced by exploiting an inherent property of a light field. One example is that an epipolar plane image (EPI), which is a 2D subset of the 4D light field, consists of many lines, and these lines have almost the same slope in a local region. This structure induces a sparse representation in the frequency domain, where most of the energy resides on a line passing through the origin. On the basis of this observation, we propose a group sparsity prior suitable for light fields to exploit their line structure fully for interpolation. Specifically, we designed the directional groups in the discrete Fourier transform (DFT) domain so that the groups can represent the concentration of the energy, and we thereby formulated an LF interpolation problem as an overlapping group lasso. We also introduce several techniques to improve the interpolation accuracy such as applying a window function, determining group weights, expanding processing blocks, and merging blocks. Our experimental results show that the proposed method can achieve better or comparable quality as compared to state-of-the-art LF interpolation methods such as convolutional neural network (CNN)-based methods.},
keywords={},
doi={10.1587/transfun.2018EAP1175},
ISSN={1745-1337},
month={January},}
Salinan
TY - JOUR
TI - Good Group Sparsity Prior for Light Field Interpolation
T2 - IEICE TRANSACTIONS on Fundamentals
SP - 346
EP - 355
AU - Shu FUJITA
AU - Keita TAKAHASHI
AU - Toshiaki FUJII
PY - 2020
DO - 10.1587/transfun.2018EAP1175
JO - IEICE TRANSACTIONS on Fundamentals
SN - 1745-1337
VL - E103-A
IS - 1
JA - IEICE TRANSACTIONS on Fundamentals
Y1 - January 2020
AB - A light field, which is equivalent to a dense set of multi-view images, has various applications such as depth estimation and 3D display. One of the essential problems in light field applications is light field interpolation, i.e., view interpolation. The interpolation accuracy is enhanced by exploiting an inherent property of a light field. One example is that an epipolar plane image (EPI), which is a 2D subset of the 4D light field, consists of many lines, and these lines have almost the same slope in a local region. This structure induces a sparse representation in the frequency domain, where most of the energy resides on a line passing through the origin. On the basis of this observation, we propose a group sparsity prior suitable for light fields to exploit their line structure fully for interpolation. Specifically, we designed the directional groups in the discrete Fourier transform (DFT) domain so that the groups can represent the concentration of the energy, and we thereby formulated an LF interpolation problem as an overlapping group lasso. We also introduce several techniques to improve the interpolation accuracy such as applying a window function, determining group weights, expanding processing blocks, and merging blocks. Our experimental results show that the proposed method can achieve better or comparable quality as compared to state-of-the-art LF interpolation methods such as convolutional neural network (CNN)-based methods.
ER -