Publication:
Perceptual Colour Features for Natural Scene Image Description and Retrieval

dc.contributor.authorSukanya Phongsuphapen_US
dc.contributor.authorKannikar Kamolraten_US
dc.contributor.otherMahidol Universityen_US
dc.contributor.otherRajabhat Universityen_US
dc.date.accessioned2018-12-11T02:40:44Z
dc.date.accessioned2019-03-14T08:04:34Z
dc.date.available2018-12-11T02:40:44Z
dc.date.available2019-03-14T08:04:34Z
dc.date.issued2016-01-12en_US
dc.description.abstract© 2015 IEEE. In this paper, we propose new image features called perceptual colour features. The features are based on twenty basic colours called emotional colours, which are used to describe the relationship between colours and emotions in psychological studies. We analyzed a colour image in L∗a∗b∗ and L∗C∗h colour spaces to link to emotions. Then, the perceptual colour features are derived from the histogram of the twenty emotional colours. We have demonstrated effectiveness of the perceptual colour features with an emotion based image retrieval application. Experiments were performed on a set of impressive natural scene images. The retrieval results were analyzed by both objective evaluation and subjective evaluation. The results revealed that the proposed perceptual colour features outperformed the conventional colour histogram features both in terms of precision and consistency with users' perception.en_US
dc.identifier.citationProceedings - 2015 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2015. (2016), 1732-1737en_US
dc.identifier.doi10.1109/SMC.2015.304en_US
dc.identifier.other2-s2.0-84964414609en_US
dc.identifier.urihttps://repository.li.mahidol.ac.th/handle/20.500.14594/43516
dc.rightsMahidol Universityen_US
dc.rights.holderSCOPUSen_US
dc.source.urihttps://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84964414609&origin=inwarden_US
dc.subjectComputer Scienceen_US
dc.subjectDecision Sciencesen_US
dc.subjectEnergyen_US
dc.subjectEngineeringen_US
dc.titlePerceptual Colour Features for Natural Scene Image Description and Retrievalen_US
dc.typeConference Paperen_US
dspace.entity.typePublication
mu.datasource.scopushttps://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84964414609&origin=inwarden_US

Files

Collections