Tampere University of Technology

TUTCRIS Research Portal

360 panorama super-resolution using deep convolutional networks

Research output: Chapter in Book/Report/Conference proceedingConference contributionScientificpeer-review

Standard

360 panorama super-resolution using deep convolutional networks. / Fakour-Sevom, Vida; Guldogan, Esin; Kämäräinen, Joni-Kristian.

VISIGRAPP 2018 - Proceedings of the 13th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications . Vol. 4 SCITEPRESS, 2018. p. 159-165.

Research output: Chapter in Book/Report/Conference proceedingConference contributionScientificpeer-review

Harvard

Fakour-Sevom, V, Guldogan, E & Kämäräinen, J-K 2018, 360 panorama super-resolution using deep convolutional networks. in VISIGRAPP 2018 - Proceedings of the 13th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications . vol. 4, SCITEPRESS, pp. 159-165, INTERNATIONAL CONFERENCE ON COMPUTER VISION THEORY AND APPLICATIONS, 1/01/00. https://doi.org/10.5220/0006618901590165

APA

Fakour-Sevom, V., Guldogan, E., & Kämäräinen, J-K. (2018). 360 panorama super-resolution using deep convolutional networks. In VISIGRAPP 2018 - Proceedings of the 13th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (Vol. 4, pp. 159-165). SCITEPRESS. https://doi.org/10.5220/0006618901590165

Vancouver

Fakour-Sevom V, Guldogan E, Kämäräinen J-K. 360 panorama super-resolution using deep convolutional networks. In VISIGRAPP 2018 - Proceedings of the 13th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications . Vol. 4. SCITEPRESS. 2018. p. 159-165 https://doi.org/10.5220/0006618901590165

Author

Fakour-Sevom, Vida ; Guldogan, Esin ; Kämäräinen, Joni-Kristian. / 360 panorama super-resolution using deep convolutional networks. VISIGRAPP 2018 - Proceedings of the 13th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications . Vol. 4 SCITEPRESS, 2018. pp. 159-165

Bibtex - Download

@inproceedings{5353c91751eb4d13ab4125fc508bcf7b,
title = "360 panorama super-resolution using deep convolutional networks",
abstract = "We propose deep convolutional neural network (CNN) based super-resolution for 360 (equirectangular) panorama images used by virtual reality (VR) display devices (e.g. VR glasses). Proposed super-resolution adopts the recent CNN architecture proposed in (Dong et al., 2016) and adapts it for equirectangular panorama images which have specific characteristics as compared to standard cameras (e.g. projection distortions). We demonstrate how adaptation can be performed by optimizing the trained network input size and fine-tuning the network parameters. In our experiments with 360 panorama images of rich natural content CNN based super-resolution achieves average PSNR improvement of 1.36 dB over the baseline (bicubic interpolation) and 1.56 dB by our equirectangular specific adaptation.",
keywords = "Deep convolutional neural network, Equirectangular panorama, Super-resolution, Virtual reality",
author = "Vida Fakour-Sevom and Esin Guldogan and Joni-Kristian K{\"a}m{\"a}r{\"a}inen",
note = "EXT={"}Guldogan, Esin{"}",
year = "2018",
doi = "10.5220/0006618901590165",
language = "English",
volume = "4",
pages = "159--165",
booktitle = "VISIGRAPP 2018 - Proceedings of the 13th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications",
publisher = "SCITEPRESS",

}

RIS (suitable for import to EndNote) - Download

TY - GEN

T1 - 360 panorama super-resolution using deep convolutional networks

AU - Fakour-Sevom, Vida

AU - Guldogan, Esin

AU - Kämäräinen, Joni-Kristian

N1 - EXT="Guldogan, Esin"

PY - 2018

Y1 - 2018

N2 - We propose deep convolutional neural network (CNN) based super-resolution for 360 (equirectangular) panorama images used by virtual reality (VR) display devices (e.g. VR glasses). Proposed super-resolution adopts the recent CNN architecture proposed in (Dong et al., 2016) and adapts it for equirectangular panorama images which have specific characteristics as compared to standard cameras (e.g. projection distortions). We demonstrate how adaptation can be performed by optimizing the trained network input size and fine-tuning the network parameters. In our experiments with 360 panorama images of rich natural content CNN based super-resolution achieves average PSNR improvement of 1.36 dB over the baseline (bicubic interpolation) and 1.56 dB by our equirectangular specific adaptation.

AB - We propose deep convolutional neural network (CNN) based super-resolution for 360 (equirectangular) panorama images used by virtual reality (VR) display devices (e.g. VR glasses). Proposed super-resolution adopts the recent CNN architecture proposed in (Dong et al., 2016) and adapts it for equirectangular panorama images which have specific characteristics as compared to standard cameras (e.g. projection distortions). We demonstrate how adaptation can be performed by optimizing the trained network input size and fine-tuning the network parameters. In our experiments with 360 panorama images of rich natural content CNN based super-resolution achieves average PSNR improvement of 1.36 dB over the baseline (bicubic interpolation) and 1.56 dB by our equirectangular specific adaptation.

KW - Deep convolutional neural network

KW - Equirectangular panorama

KW - Super-resolution

KW - Virtual reality

U2 - 10.5220/0006618901590165

DO - 10.5220/0006618901590165

M3 - Conference contribution

VL - 4

SP - 159

EP - 165

BT - VISIGRAPP 2018 - Proceedings of the 13th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications

PB - SCITEPRESS

ER -