Publicação
The role of background colour in pollen recognition task using CNN
| Resumo: | Pollen recognition is a crucial but challenging task in addressing a variety of questions like pollination or palaeobotany, but also for other fields of research, e.g., allergology, melissopalynology or forensics. State-of-the-art methods mainly use deep learning CNNs for pollen recognition, however, we observe that existing published approaches use original images without study the possible biased recognition due to pollen’s background colour. In this paper, we evaluate the DenseNet model trained with original images and with segmented images (remove background) and analyse network’s predictive performance under these conditions using a cross evaluation approach. An accuracy of 97.4% was achieved that represents one of the best successes rate when weighted with the number of taxa of any attempt at automated pollen analysis currently documented in the literature. From these results, we confirm the existence of background specific influence in the recognition task. |
|---|---|
| Autores principais: | Monteiro, Fernando C. |
| Outros Autores: | Pinto, Cristina M.; Rufino, José |
| Assunto: | Pollen recognition Deep learning Convolutional neural network |
| Ano: | 2021 |
| País: | Portugal |
| Tipo de documento: | documento de conferência |
| Tipo de acesso: | acesso aberto |
| Instituição associada: | Universidade Politécnica de Bragança |
| Idioma: | inglês |
| Origem: | Biblioteca Digital da UPB |
| _version_ | 1877782666239016960 |
|---|---|
| author | Monteiro, Fernando C. |
| author2 | Pinto, Cristina M. Rufino, José |
| author2_role | author author |
| author_facet | Monteiro, Fernando C. Pinto, Cristina M. Rufino, José |
| author_role | author |
| contributor_name_str_mv | Biblioteca Digital da UPB |
| country_str | PT |
| creators_json_txt | [{\"Person.name\":\"Monteiro, Fernando C.\",\"Person.identifier.cienciaID\":\"2019-BDBF-10E2\",\"Person.identifier.orcid\":\"0000-0002-1421-8006\",\"Person.identifier.researcherID\":\"H-9213-2016\",\"Person.identifier.scopusAuthorID\":\"8986162600\"},{\"Person.name\":\"Pinto, Cristina M.\"},{\"Person.name\":\"Rufino, José\",\"Person.identifier.cienciaID\":\"C414-F47F-6323\",\"Person.identifier.orcid\":\"0000-0002-1344-8264\",\"Person.identifier.scopusAuthorID\":\"55947199100\",\"Person.identifier.scopusAuthorID\":\"57188967176\"}] |
| datacite.contributors.contributor.contributorName.fl_str_mv | Biblioteca Digital da UPB |
| datacite.creators.creator.creatorName.fl_str_mv | Monteiro, Fernando C. Pinto, Cristina M. Rufino, José |
| datacite.date.Accepted.fl_str_mv | 2021-01-01T00:00:00Z |
| datacite.date.available.fl_str_mv | 2022-01-14T09:31:13Z |
| datacite.date.embargoed.fl_str_mv | 2022-01-14T09:31:13Z |
| datacite.rights.fl_str_mv | http://purl.org/coar/access_right/c_abf2 |
| datacite.subjects.subject.fl_str_mv | Pollen recognition Deep learning Convolutional neural network |
| datacite.titles.title.fl_str_mv | The role of background colour in pollen recognition task using CNN |
| dc.contributor.none.fl_str_mv | Biblioteca Digital da UPB |
| dc.creator.none.fl_str_mv | Monteiro, Fernando C. Pinto, Cristina M. Rufino, José |
| dc.date.Accepted.fl_str_mv | 2021-01-01T00:00:00Z |
| dc.date.available.fl_str_mv | 2022-01-14T09:31:13Z |
| dc.date.embargoed.fl_str_mv | 2022-01-14T09:31:13Z |
| dc.format.none.fl_str_mv | application/pdf |
| dc.identifier.none.fl_str_mv | http://hdl.handle.net/10198/24643 |
| dc.language.none.fl_str_mv | eng |
| dc.publisher.none.fl_str_mv | American Council on Science & Education |
| dc.rights.cclincense.fl_str_mv | http://creativecommons.org/licenses/by/4.0/ |
| dc.rights.none.fl_str_mv | http://purl.org/coar/access_right/c_abf2 |
| dc.subject.none.fl_str_mv | Pollen recognition Deep learning Convolutional neural network |
| dc.title.fl_str_mv | The role of background colour in pollen recognition task using CNN |
| dc.type.none.fl_str_mv | http://purl.org/coar/resource_type/c_c94f |
| description | Pollen recognition is a crucial but challenging task in addressing a variety of questions like pollination or palaeobotany, but also for other fields of research, e.g., allergology, melissopalynology or forensics. State-of-the-art methods mainly use deep learning CNNs for pollen recognition, however, we observe that existing published approaches use original images without study the possible biased recognition due to pollen’s background colour. In this paper, we evaluate the DenseNet model trained with original images and with segmented images (remove background) and analyse network’s predictive performance under these conditions using a cross evaluation approach. An accuracy of 97.4% was achieved that represents one of the best successes rate when weighted with the number of taxa of any attempt at automated pollen analysis currently documented in the literature. From these results, we confirm the existence of background specific influence in the recognition task. |
| dirty | 0 |
| eu_rights_str_mv | openAccess |
| format | conferenceObject |
| fulltext.url.fl_str_mv | https://bibliotecadigital.unipb.pt/bitstreams/ecc29adf-39c6-4b4e-b123-92f921922f89/download |
| id | ipb_f21c82ee345f1fc6fb4e373c8bfc2afc |
| identifier.url.fl_str_mv | http://hdl.handle.net/10198/24643 |
| inst_facet_str | urn:organizationAcronym:unipb{{{_:::_}}}Universidade Politécnica de Bragança |
| instacron_str | unipb |
| institution | Universidade Politécnica de Bragança |
| instname_str | Universidade Politécnica de Bragança |
| language | eng |
| network_acronym_str | ipb |
| network_name_str | Biblioteca Digital da UPB |
| oai_identifier_str | oai:bibliotecadigital.unipb.pt:10198/24643 |
| organization_str_mv | urn:organizationAcronym:unipb |
| person_str_mv | Monteiro, Fernando C. Monteiro, Fernando C. https://www.ciencia-id.pt/2019-BDBF-10E2 2019-BDBF-10E2 http://orcid.org/0000-0002-1421-8006 0000-0002-1421-8006 Pinto, Cristina M. Rufino, José Rufino, José https://www.ciencia-id.pt/C414-F47F-6323 C414-F47F-6323 http://orcid.org/0000-0002-1344-8264 0000-0002-1344-8264 |
| publishDate | 2021 |
| publisher.none.fl_str_mv | American Council on Science & Education |
| repo_facet_str | urn:repositoryAcronym:ipb{{{_:::_}}}Biblioteca Digital da UPB |
| reponame_str | Biblioteca Digital da UPB |
| repository_id_str | urn:repositoryAcronym:ipb |
| service_str_mv | urn:repositoryAcronym:ipb |
| spelling | engAmerican Council on Science & Educationpt_PTPollen recognition is a crucial but challenging task in addressing a variety of questions like pollination or palaeobotany, but also for other fields of research, e.g., allergology, melissopalynology or forensics. State-of-the-art methods mainly use deep learning CNNs for pollen recognition, however, we observe that existing published approaches use original images without study the possible biased recognition due to pollen’s background colour. In this paper, we evaluate the DenseNet model trained with original images and with segmented images (remove background) and analyse network’s predictive performance under these conditions using a cross evaluation approach. An accuracy of 97.4% was achieved that represents one of the best successes rate when weighted with the number of taxa of any attempt at automated pollen analysis currently documented in the literature. From these results, we confirm the existence of background specific influence in the recognition task.application/pdfpt_PTThe role of background colour in pollen recognition task using CNNPersonalMonteiro, Fernando C.DSpacehttp://dspace.org/items/363b6c37-282c-4cd6-bb54-3c97cc700d78DSpacehttp://dspace.org/items/363b6c37-282c-4cd6-bb54-3c97cc700d78MonteiroFernando C.Ciência IDhttps://www.ciencia-id.pt2019-BDBF-10E2ORCIDhttp://orcid.org0000-0002-1421-8006Researcher IDhttps://www.researcherid.comH-9213-2016Scopus Author IDhttps://www.scopus.com8986162600Pinto, Cristina M.PersonalRufino, JoséDSpacehttp://dspace.org/items/1e24d2ce-a354-442a-bef8-eebadd94b385DSpacehttp://dspace.org/items/1e24d2ce-a354-442a-bef8-eebadd94b385RufinoJoséCiência IDhttps://www.ciencia-id.ptC414-F47F-6323ORCIDhttp://orcid.org0000-0002-1344-8264Scopus Author IDhttps://www.scopus.com55947199100Scopus Author IDhttps://www.scopus.com57188967176HostingInstitutionOrganizationalBiblioteca Digital da UPBe-mailmailto:dspace@unipb.ptdspace@unipb.pt2022-01-14T09:31:13Z20212021-01-01T00:00:00ZHandlehttp://hdl.handle.net/10198/24643http://purl.org/coar/access_right/c_abf2open accessPollen recognitionDeep learningConvolutional neural network326346 bytesother research producthttp://purl.org/coar/resource_type/c_c94fconference object2021http://creativecommons.org/licenses/by/4.0/http://purl.org/coar/access_right/c_abf2application/pdffulltexthttps://bibliotecadigital.unipb.pt/bitstreams/ecc29adf-39c6-4b4e-b123-92f921922f89/downloadThe 2021 World Congress in Computer Science, Computer Engineering, and Applied Computing (CSCE 2021)88Las Vegas - USA |
| spellingShingle | The role of background colour in pollen recognition task using CNN Monteiro, Fernando C. Pollen recognition Deep learning Convolutional neural network |
| status | SINGLETON |
| subject.fl_str_mv | Pollen recognition Deep learning Convolutional neural network |
| title | The role of background colour in pollen recognition task using CNN |
| title_full | The role of background colour in pollen recognition task using CNN |
| title_fullStr | The role of background colour in pollen recognition task using CNN |
| title_full_unstemmed | The role of background colour in pollen recognition task using CNN |
| title_short | The role of background colour in pollen recognition task using CNN |
| title_sort | The role of background colour in pollen recognition task using CNN |
| topic | Pollen recognition Deep learning Convolutional neural network |
| topic_facet | Pollen recognition Deep learning Convolutional neural network |
| url | http://hdl.handle.net/10198/24643 |
| visible | 1 |
Atividades financiadas
Não foram encontrados projetos financiados.