Show simple item record

dc.contributor.authorLi, CY
dc.contributor.authorMazzon, R
dc.contributor.authorCavallaro, A
dc.date.accessioned2019-01-31T11:14:35Z
dc.date.available2019-01-31T11:14:35Z
dc.date.issued2019-01-01
dc.identifier.citationLi, C., Mazzon, R. and Cavallaro, A. (2018). An Online Platform for Underwater Image Quality Evaluation. Pattern Recognition and Information Forensics, [online] pp.37-44. Available at: https://link.springer.com/chapter/10.1007%2F978-3-030-05792-3_4 [Accessed 31 Jan. 2019].en_US
dc.identifier.isbn9783030057916
dc.identifier.issn0302-9743
dc.identifier.urihttps://qmro.qmul.ac.uk/xmlui/handle/123456789/55048
dc.description.abstractWith the miniaturisation of underwater cameras, the volume of available underwater images has been considerably increasing. However, underwater images are degraded by the absorption and scattering of light in water. Image processing methods exist that aim to compensate for these degradations, but there are no standard quality evaluation measures or testing datasets for a systematic empirical comparison. For this reason, we propose PUIQE, an online platform for underwater image quality evaluation, which is inspired by other computer vision areas whose progress has been accelerated by evaluation platforms. PUIQE supports the comparison of methods through standard datasets and objective evaluation measures: quality scores for images uploaded on the platform are automatically computed and published in a leaderboard, which enables the ranking of methods. We hope that PUIQE will stimulate and facilitate the development of underwater image processing algorithms to improve underwater images.en_US
dc.format.extent37 - 44
dc.publisherSpringeren_US
dc.titleAn Online Platform for Underwater Image Quality Evaluationen_US
dc.typeConference Proceedingen_US
dc.rights.holder© Springer Nature Switzerland AG 2019
dc.identifier.doi10.1007/978-3-030-05792-3_4
pubs.notesNot knownen_US
pubs.publication-statusPublisheden_US
pubs.volume11188 LNCSen_US
rioxxterms.funderDefault funderen_US
rioxxterms.identifier.projectDefault projecten_US


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record