Show simple item record  

dc.contributor.authorMo, Jeffen_NZ
dc.contributor.authorFrank, Eibeen_NZ
dc.contributor.authorVetrova, Varvaraen_NZ
dc.contributor.editorPeng, Weien_NZ
dc.contributor.editorAlahakoon, Dammindaen_NZ
dc.contributor.editorLi, Xiaodongen_NZ
dc.coverage.spatialMelbourne, Australiaen_NZ
dc.date.accessioned2017-08-09T21:15:29Z
dc.date.available2017en_NZ
dc.date.available2017-08-09T21:15:29Z
dc.date.issued2017en_NZ
dc.identifier.citationMo, J., Frank, E., & Vetrova, V. (2017). Large-scale automatic species identification. In W. Peng, D. Alahakoon, & X. Li (Eds.), Proceedings of 30th Australasian Joint Conference on Advances in Artificial Intelligence (Vol. LNCS 10400, pp. 301–312). Springer, Cham: Springer. https://doi.org/10.1007/978-3-319-63004-5_24en
dc.identifier.isbn978-3-319-63004-5en_NZ
dc.identifier.isbn978-3-319-63003-8
dc.identifier.urihttps://hdl.handle.net/10289/11266
dc.description.abstractThe crowd-sourced Naturewatch GBIF dataset is used to obtain a species classification dataset containing approximately 1.2 million photos of nearly 20 thousand different species of biological organisms observed in their natural habitat. We present a general hierarchical species identification system based on deep convolutional neural networks trained on the NatureWatch dataset. The dataset contains images taken under a wide variety of conditions and is heavily imbalanced, with most species associated with only few images. We apply multi-view classification as a way to lend more influence to high frequency details, hierarchical fine-tuning to help with class imbalance and provide regularisation, and automatic specificity control for optimising classification depth. Our system achieves 55.8% accuracy when identifying individual species and around 90% accuracy at an average taxonomy depth of 5.1—equivalent to the taxonomic rank of “family”—when applying automatic specificity control.
dc.format.mimetypeapplication/pdf
dc.language.isoen
dc.publisherSpringeren_NZ
dc.rights© Springer International Publishing AG 2017.This is the author's accepted version. The final publication is available at Springer via https://doi.org/10.1007/978-3-319-63004-5_24
dc.sourceAI 2017: Advances in Artificial Intelligenceen_NZ
dc.subjectcomputer scienceen_NZ
dc.subjectspecies identificationen_NZ
dc.subjectconvolutional neural networksen_NZ
dc.subjectMachine learning
dc.titleLarge-scale automatic species identificationen_NZ
dc.typeConference Contribution
dc.identifier.doi10.1007/978-3-319-63004-5_24en_NZ
dc.relation.isPartOfProceedings of 30th Australasian Joint Conference on Advances in Artificial Intelligenceen_NZ
pubs.begin-page301
pubs.declined2017-06-13T15:06:47.926+1200
pubs.elements-id194358
pubs.end-page312
pubs.finish-date2017-08-20en_NZ
pubs.place-of-publicationSpringer, Cham
pubs.publication-statusPublisheden_NZ
pubs.start-date2017-08-19en_NZ
pubs.volumeLNCS 10400en_NZ


Files in this item

This item appears in the following Collection(s)

Show simple item record