Light field denoising via anisotropic parallax analysis in a CNN framework

Light field (LF) cameras provide perspective information of scenes by taking directional measurements of the focusing light rays. The raw outputs are usually dark with additive camera noise, which impedes subsequent processing and applications. We propose a novel LF denoising framework based on anis...

Full description

Saved in:
Bibliographic Details
Main Authors: Chen, Jie, Hou, Junhui, Chau, Lap-Pui
Other Authors: School of Electrical and Electronic Engineering
Format: Article
Language:English
Published: 2020
Subjects:
Online Access:https://hdl.handle.net/10356/142560
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
id sg-ntu-dr.10356-142560
record_format dspace
spelling sg-ntu-dr.10356-1425602020-06-24T06:11:56Z Light field denoising via anisotropic parallax analysis in a CNN framework Chen, Jie Hou, Junhui Chau, Lap-Pui School of Electrical and Electronic Engineering Engineering::Electrical and electronic engineering Anisotropic Parallax Feature Convolutional Neural Networks (CNN) Light field (LF) cameras provide perspective information of scenes by taking directional measurements of the focusing light rays. The raw outputs are usually dark with additive camera noise, which impedes subsequent processing and applications. We propose a novel LF denoising framework based on anisotropic parallax analysis (APA). Two convolutional neural networks are jointly designed for the task: first, the structural parallax synthesis network predicts the parallax details for the entire LF based on a set of anisotropic parallax features. These novel features can efficiently capture the high-frequency perspective components of a LF from noisy observations. Second, the view-dependent detail compensation network restores non-Lambertian variation to each LF view by involving view-specific spatial energies. Extensive experiments show that the proposed APA LF denoiser provides a much better denoising performance than state-of-the-art methods in terms of visual quality and in preservation of parallax details. NRF (Natl Research Foundation, S’pore) 2020-06-24T06:11:56Z 2020-06-24T06:11:56Z 2018 Journal Article Chen, J., Hou, J., & Chau, L.-P. (2018). Light field denoising via anisotropic parallax analysis in a CNN framework. IEEE Signal Processing Letters, 25(9), 1403-1407. doi:10.1109/LSP.2018.2861212 1070-9908 https://hdl.handle.net/10356/142560 10.1109/LSP.2018.2861212 2-s2.0-85050767622 9 25 1403 1407 en IEEE Signal Processing Letters © 2018 IEEE. All rights reserved.
institution Nanyang Technological University
building NTU Library
country Singapore
collection DR-NTU
language English
topic Engineering::Electrical and electronic engineering
Anisotropic Parallax Feature
Convolutional Neural Networks (CNN)
spellingShingle Engineering::Electrical and electronic engineering
Anisotropic Parallax Feature
Convolutional Neural Networks (CNN)
Chen, Jie
Hou, Junhui
Chau, Lap-Pui
Light field denoising via anisotropic parallax analysis in a CNN framework
description Light field (LF) cameras provide perspective information of scenes by taking directional measurements of the focusing light rays. The raw outputs are usually dark with additive camera noise, which impedes subsequent processing and applications. We propose a novel LF denoising framework based on anisotropic parallax analysis (APA). Two convolutional neural networks are jointly designed for the task: first, the structural parallax synthesis network predicts the parallax details for the entire LF based on a set of anisotropic parallax features. These novel features can efficiently capture the high-frequency perspective components of a LF from noisy observations. Second, the view-dependent detail compensation network restores non-Lambertian variation to each LF view by involving view-specific spatial energies. Extensive experiments show that the proposed APA LF denoiser provides a much better denoising performance than state-of-the-art methods in terms of visual quality and in preservation of parallax details.
author2 School of Electrical and Electronic Engineering
author_facet School of Electrical and Electronic Engineering
Chen, Jie
Hou, Junhui
Chau, Lap-Pui
format Article
author Chen, Jie
Hou, Junhui
Chau, Lap-Pui
author_sort Chen, Jie
title Light field denoising via anisotropic parallax analysis in a CNN framework
title_short Light field denoising via anisotropic parallax analysis in a CNN framework
title_full Light field denoising via anisotropic parallax analysis in a CNN framework
title_fullStr Light field denoising via anisotropic parallax analysis in a CNN framework
title_full_unstemmed Light field denoising via anisotropic parallax analysis in a CNN framework
title_sort light field denoising via anisotropic parallax analysis in a cnn framework
publishDate 2020
url https://hdl.handle.net/10356/142560
_version_ 1681057487599435776