TUTCRIS - Tampereen teknillinen yliopisto

TUTCRIS

Deep multiresolution color constancy

Tutkimustuotosvertaisarvioitu

Standard

Deep multiresolution color constancy. / Aytekin, Caglar; Nikkanen, Jarno; Gabbouj, Moncef.

2017 IEEE International Conference on Image Processing, ICIP 2017 - Proceedings. IEEE COMPUTER SOCIETY PRESS, 2018. s. 3735-3739.

Tutkimustuotosvertaisarvioitu

Harvard

Aytekin, C, Nikkanen, J & Gabbouj, M 2018, Deep multiresolution color constancy. julkaisussa 2017 IEEE International Conference on Image Processing, ICIP 2017 - Proceedings. IEEE COMPUTER SOCIETY PRESS, Sivut 3735-3739, IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING, 1/01/00. https://doi.org/10.1109/ICIP.2017.8296980

APA

Aytekin, C., Nikkanen, J., & Gabbouj, M. (2018). Deep multiresolution color constancy. teoksessa 2017 IEEE International Conference on Image Processing, ICIP 2017 - Proceedings (Sivut 3735-3739). IEEE COMPUTER SOCIETY PRESS. https://doi.org/10.1109/ICIP.2017.8296980

Vancouver

Aytekin C, Nikkanen J, Gabbouj M. Deep multiresolution color constancy. julkaisussa 2017 IEEE International Conference on Image Processing, ICIP 2017 - Proceedings. IEEE COMPUTER SOCIETY PRESS. 2018. s. 3735-3739 https://doi.org/10.1109/ICIP.2017.8296980

Author

Aytekin, Caglar ; Nikkanen, Jarno ; Gabbouj, Moncef. / Deep multiresolution color constancy. 2017 IEEE International Conference on Image Processing, ICIP 2017 - Proceedings. IEEE COMPUTER SOCIETY PRESS, 2018. Sivut 3735-3739

Bibtex - Lataa

@inproceedings{f27a9a72c86a4b61ba663812f13a0f5d,
title = "Deep multiresolution color constancy",
abstract = "In this paper, a computational color constancy method is proposed via estimating the illuminant chromaticity in a scene by pooling from many local estimates. To this end, first, for each image in a dataset, we form an image pyramid consisting of several scales of the original image. Next, local patches of certain size are extracted from each scale in this image pyramid. Then, a convolutional neural network is trained to estimate the illuminant chromaticity per-patch. Finally, two more consecutive trainings are conducted, where the estimation is made per-image via taking the mean (1st training) and median (2nd training) of local estimates. The proposed method is shown to outperform the state-of-the-art in a widely used color constancy dataset.",
keywords = "Color constancy, Deep learning, Illuminant chromaticity estimation, Local estimation, Multi-resolution",
author = "Caglar Aytekin and Jarno Nikkanen and Moncef Gabbouj",
note = "jufoid=57423",
year = "2018",
month = "2",
day = "20",
doi = "10.1109/ICIP.2017.8296980",
language = "English",
publisher = "IEEE COMPUTER SOCIETY PRESS",
pages = "3735--3739",
booktitle = "2017 IEEE International Conference on Image Processing, ICIP 2017 - Proceedings",

}

RIS (suitable for import to EndNote) - Lataa

TY - GEN

T1 - Deep multiresolution color constancy

AU - Aytekin, Caglar

AU - Nikkanen, Jarno

AU - Gabbouj, Moncef

N1 - jufoid=57423

PY - 2018/2/20

Y1 - 2018/2/20

N2 - In this paper, a computational color constancy method is proposed via estimating the illuminant chromaticity in a scene by pooling from many local estimates. To this end, first, for each image in a dataset, we form an image pyramid consisting of several scales of the original image. Next, local patches of certain size are extracted from each scale in this image pyramid. Then, a convolutional neural network is trained to estimate the illuminant chromaticity per-patch. Finally, two more consecutive trainings are conducted, where the estimation is made per-image via taking the mean (1st training) and median (2nd training) of local estimates. The proposed method is shown to outperform the state-of-the-art in a widely used color constancy dataset.

AB - In this paper, a computational color constancy method is proposed via estimating the illuminant chromaticity in a scene by pooling from many local estimates. To this end, first, for each image in a dataset, we form an image pyramid consisting of several scales of the original image. Next, local patches of certain size are extracted from each scale in this image pyramid. Then, a convolutional neural network is trained to estimate the illuminant chromaticity per-patch. Finally, two more consecutive trainings are conducted, where the estimation is made per-image via taking the mean (1st training) and median (2nd training) of local estimates. The proposed method is shown to outperform the state-of-the-art in a widely used color constancy dataset.

KW - Color constancy

KW - Deep learning

KW - Illuminant chromaticity estimation

KW - Local estimation

KW - Multi-resolution

U2 - 10.1109/ICIP.2017.8296980

DO - 10.1109/ICIP.2017.8296980

M3 - Conference contribution

SP - 3735

EP - 3739

BT - 2017 IEEE International Conference on Image Processing, ICIP 2017 - Proceedings

PB - IEEE COMPUTER SOCIETY PRESS

ER -